def __init__(self, model=None):
Solver.__init__(self)
self.problem = Cplex()
self.status_mapping = {
self.problem.solution.status.optimal:
Status.OPTIMAL,
self.problem.solution.status.optimal_tolerance:
Status.OPTIMAL,
self.problem.solution.status.unbounded:
Status.UNBOUNDED,
self.problem.solution.status.infeasible:
Status.INFEASIBLE,
self.problem.solution.status.infeasible_or_unbounded:
Status.INF_OR_UNB,
self.problem.solution.status.MIP_optimal:
Status.OPTIMAL,
self.problem.solution.status.MIP_unbounded:
Status.UNBOUNDED,
self.problem.solution.status.MIP_infeasible:
Status.INFEASIBLE,
self.problem.solution.status.MIP_infeasible_or_unbounded:
Status.INF_OR_UNB
}
self.vartype_mapping = {
VarType.BINARY: self.problem.variables.type.binary,
VarType.INTEGER: self.problem.variables.type.integer,
VarType.CONTINUOUS: self.problem.variables.type.continuous
}
self.parameter_mapping = {
Parameter.TIME_LIMIT: self.problem.parameters.timelimit,
Parameter.FEASIBILITY_TOL:
self.problem.parameters.simplex.tolerances.feasibility,
Parameter.OPTIMALITY_TOL:
self.problem.parameters.simplex.tolerances.optimality,
Parameter.INT_FEASIBILITY_TOL:
self.problem.parameters.mip.tolerances.integrality,
Parameter.MIP_ABS_GAP:
self.problem.parameters.mip.tolerances.mipgap,
Parameter.MIP_REL_GAP:
self.problem.parameters.mip.tolerances.absmipgap,
Parameter.POOL_SIZE: self.problem.parameters.mip.limits.populate,
Parameter.POOL_GAP: self.problem.parameters.mip.pool.relgap
}
self.set_parameters(default_parameters)
self.set_logging(False)
self._cached_lin_obj = {}
self._cached_sense = None
self._cached_lower_bounds = {}
self._cached_upper_bounds = {}
self._cached_vars = []
self._cached_constrs = []
if model:
self.build_problem(model)
def solve(problem: cplex.Cplex, dataset: Dataset):
global best_solution, best_score
# Ищем решение с текущими ограничениями
problem.solve()
# Если решение не удовлетворяет ограничениям, не идем в эту ветку
if (problem.solution.get_status() is not 1):
print("Broken model")
return None
obj = problem.solution.get_objective_value()
# Если решение хуже лучшего, не идем в эту ветку
if (obj <= best_score):
return None
prev_ans = obj
add_constr = lambda constraint: problem.linear_constraints.add(
lin_expr=
[cplex.SparsePair(constraint, val=[1.0] * len(constraint))],
rhs=[1.0],
names=["c_" + str(np.random.randint(low=10, high=1000000))],
senses=['L'])
while True:
constraint = BranchAndCut.separation(dataset,
problem.solution.get_values())
if (len(constraint) <= 1):
break
add_constr(constraint)
problem.solve()
obj = problem.solution.get_objective_value()
if (obj < best_score):
return None
elif np.isclose(prev_ans, obj, atol=10e-6):
break
prev_ans = obj
solution = problem.solution.get_values()
if BranchAndCut.validate_integer_solution(solution):
constraints = BranchAndCut.check_solution(problem, dataset)
if len(constraints) == 0:
if obj > best_score:
best_solution = solution
best_score = obj
print("New best result: ", best_score)
else:
for constr in constraints:
add_constr(constr)
BranchAndCut.solve(problem, dataset)
return
branching_var = BranchAndCut.get_branching_var(solution)
branching_var = [int(branching_var)]
for val in [1.0, 0.0]:
constr = problem.linear_constraints.add(
lin_expr=[cplex.SparsePair(branching_var, val=[1.0])],
rhs=[val],
names=["e_" + str(np.random.randint(low=10, high=1000000))],
senses=['E'])
BranchAndCut.solve(problem, dataset)
problem.linear_constraints.delete(constr)
def __init__(self, cplex=None):
if cplex:
self._model = Cplex(cplex._model)
else:
self._model = Cplex()
self._init_lin()
# to avoid a variable with index 0
self._model.variables.add(
names=['_dummy_'], types=[self._model.variables.type.continuous])
self._var_id = {'_dummy_': 0}
def __init__(self, problem):
#instance variable: the solution set found by solver
self.cplexSolutionSet = []
#instance variable: the solution map, the key is the solution obj values, the value is the solution
self.cplexResultMap = {}
#instance variable: the map of the solutions in the pareto front
self.cplexParetoSet = {}
# problem
self.problem = problem
# solver
self.solver = Cplex()
def create_cplex(cls, verbose=False):
cpx = Cplex()
# see if we want it to be verbose sometimes
if not verbose:
cpx.set_log_stream(None)
cpx.set_results_stream(None)
cpx.set_error_stream(None)
cpx.set_warning_stream(None)
# disable datacheck
cpx_datacheck_id = 1056
setintparam(cpx._env._e, cpx_datacheck_id, 0)
return cpx
def __init__(self, cplex=None):
try:
if cplex:
self._model = Cplex(cplex._model)
else:
self._model = Cplex()
except NameError:
raise NameError('CPLEX is not installed. See https://www.ibm.com/support/knowledgecenter/SSSA5P_12.8.0/ilog.odms.studio.help/Optimization_Studio/topics/COS_home.html')
self._init_lin()
# to avoid a variable with index 0
self._model.variables.add(names=['_dummy_'], types=[self._model.variables.type.continuous])
self._var_id = {'_dummy_': 0}
def set_cplex_objective(cpx: cplex.Cplex,
c,
Q=None,
epsilon: float = 1e-4) -> cplex.Cplex:
"""
set_cplex_objective(cpx, c, Q) sets the objective of cplex object cpx
to have linear objective coefficients c
and quadratic objective coefficients Q
essentially sets problem to have objective of the form
min cTx + xT Q x
with original constraints etc
:param cpx: cplex object to modify
:param c: linear objective coefficients
:param Q: quadratic objective coefficients
:return: cplex object with set objective coefficients
"""
n = cpx.variables.get_num()
assert len(c) == n, "c must have {} items but len(c) is {}".format(
n, len(c))
# set linear coefficients
# cTx
for i in range(n):
cpx.objective.set_linear(i, float(c[i]))
if Q is not None and len(Q.nonzero()[0]) > 0:
assert len(Q) == n, "Q must have {} items but len(Q) is {}".format(
n, len(Q))
# quadratic coefficients
# x.T * Q * x
quadratic_coefs = []
for i in range(Q.shape[0]):
indices = Q[i].nonzero()[0].tolist()
if len(indices) == 0:
sparse_pair = cplex.SparsePair(ind=[], val=[])
else:
vals = Q[i, indices].tolist()
sparse_pair = cplex.SparsePair(ind=indices, val=vals)
quadratic_coefs.append(sparse_pair)
cpx.objective.set_quadratic(quadratic_coefs)
else:
cpx.set_problem_type(cpx.problem_type.MILP)
cpx.objective.set_sense(cpx.objective.sense.minimize)
cpx.cleanup(epsilon=epsilon)
return cpx
def copy_cplex(cpx):
cpx_copy = Cplex(cpx)
cpx_parameters = cpx.parameters.get_changed()
for (pname, pvalue) in cpx_parameters:
phandle = reduce(getattr, str(pname).split("."), cpx_copy)
phandle.set(pvalue)
return cpx_copy
def __init__(self, cplex=None):
if not _HAS_CPLEX:
raise MissingOptionalLibraryError(
libname='CPLEX',
name='SimpleCPLEX',
pip_install='pip install qiskit-aqua[cplex]')
if cplex:
self._model = Cplex(cplex._model)
else:
self._model = Cplex()
self._init_lin()
# to avoid a variable with index 0
self._model.variables.add(names=['_dummy_'], types=[self._model.variables.type.continuous])
self._var_id = {'_dummy_': 0}
def branching(self, problem: cplex.Cplex):
try:
problem.solve()
solution_values = problem.solution.get_values()
except cplex.exceptions.CplexSolverError:
return list()
if sum(solution_values) > self.current_max_clique_size:
branching_variable = self.get_branching_variable(solution_values)
if branching_variable is None:
current_clique = list(index for index, value in enumerate(solution_values) if value == 1.0)
self.current_max_clique_size = len(current_clique) if self.current_max_clique_size \
< len(current_clique)else self.current_max_clique_size
return current_clique
return max(self.branching(self.add_constraint(cplex.Cplex(problem), branching_variable, 1.0)),
self.branching(self.add_constraint(cplex.Cplex(problem), branching_variable, 0.0)),
key=lambda list: len(list))
return list()
def __init__(self, A):
self.upper_bound = 1000
self.eps = 1e-10
self.dimension = 0
try:
self.cpl = Cplex()
except NameError, CplexSolverError:
raise CplexNotInstalledError()
def lp(a, b, c):
# number of equations (m) and unknowns (n)
m, n = a.shape
# objective function and lower bounds
obj = c.copy()
lb = zeros(n)
# constraints
count = 0
sense = ""
rows = []
cols = []
vals = []
rhs = []
for i in range(m):
rows.extend([count for k in range(n)])
cols.extend([k for k in range(n)])
vals.extend(a[i, :])
rhs.append(b[i])
sense += "L"
count += 1
# cplex problem variable
prob = Cplex()
# quiet results
#prob.set_results_stream(None)
# maximiation problem
prob.objective.set_sense(prob.objective.sense.maximize)
# problem variables
prob.variables.add(obj=obj, lb=lb)
#for j in range(prob.variables.get_num()):
# prob.variables.set_types(j,prob.variables.type.integer)
# linear constraints
prob.linear_constraints.add(rhs=rhs, senses=sense)
prob.linear_constraints.set_coefficients(zip(rows, cols, vals))
# alg method
alg = prob.parameters.lpmethod.values
prob.parameters.lpmethod.set(alg.auto)
# solve problem
prob.solve()
# solution variables
var = prob.solution.get_values()
x = var[0:n]
opt = prob.solution.get_objective_value()
# return
return opt, x
def modelisation():
# Variables
m, L, li, qi = read_instance_file(argv[1])
# Modèle pour résoudre le problème de minimisation des planches utilisées
model = Cplex()
model.set_results_stream(None)
# Variables de décision du modèle
model_variables = list(range(len(li)))
model.variables.add(obj=[1 for j in model_variables])
# Contraintes du modèle
model_contraintes = range(len(qi))
model.linear_constraints.add(
lin_expr=[SparsePair() for j in model_contraintes],
senses=["G" for j in model_contraintes],
rhs=qi)
for var_index in model_variables:
model.linear_constraints.set_coefficients(var_index, var_index,
int(L / li[var_index]))
# Modèle utilisé pour générer des pattern en utilisant
# la méthode Column generation de Gilmore-Gomory
pattern_model = Cplex()
pattern_model.set_results_stream(None)
# Variable de décision
panneaux_indices = range(len(li))
pattern_model.variables.add(
types=[pattern_model.variables.type.integer for j in panneaux_indices])
pattern_model.variables.add(obj=[1], lb=[1], ub=[1])
# L'unique contrainte ici est que la taille total des panneaux ne peut être
# plus grande que la longueur de la planche L.
pattern_model.linear_constraints.add(
lin_expr=[SparsePair(ind=panneaux_indices, val=li)],
senses=["L"],
rhs=[L])
# Définir l'objectif (Minimisation)
pattern_model.objective.set_sense(pattern_model.objective.sense.minimize)
return m, model, pattern_model, model_contraintes, model_variables, panneaux_indices
def formulate(self):
self.set_aux()
self.M = Cplex() if self.solver == 'cplex' else Model()
self.addVars()
self.addCons()
self.setObj()
self.p, self.a, self.obj, self.converged, self.tot_its = None,None,None,False,0
if self.verbose:
print 'solving with', self.solver
print '%2s %10s %8s %4s %3s %10s' % ('it', 'worst line', 'viol', 'stat',
'its', 'obj')
def select_winners(ads, max_duration):
'''
Computes the set of winners that has the maximum total price.
The numbers of winners are returned.
'''
from cplex import Cplex
prob = Cplex()
prob.variables.add([ad['bid'] for ad in ads], types=prob.variables.type.binary * len(ads))
prob.linear_constraints.add([[range(len(ads)), [ad['duration'] for ad in ads]]], 'L', [max_duration])
prob.objective.set_sense(prob.objective.sense.maximize)
prob.solve()
winners = set()
losers = set()
sol = prob.solution.get_values()
print sol
for i, ad in enumerate(ads):
if sol[i] > 0.5:
winners.add(ad)
else:
losers.add(ad)
return winners, losers
def copy_cplex(cpx):
"""
Copy a Cplex object
:param cpx: Cplex object
:return: Copy of Cplex object
"""
cpx_copy = Cplex(cpx)
cpx_parameters = cpx.parameters.get_changed()
for (pname, pvalue) in cpx_parameters:
phandle = reduce(getattr, str(pname).split("."), cpx_copy)
phandle.set(pvalue)
return cpx_copy
def branch_and_bound(model: cplex.Cplex, optimal_objective_value, optimal_values):
if check_timeout():
global is_timeout
is_timeout = True
return optimal_objective_value, optimal_values
model.solve()
new_result = model.solution.get_objective_value()
new_variables = model.solution.get_values()
integer_new_result = round_with_eps(new_result)
if not is_result_improved(optimal_objective_value, integer_new_result):
# print("BOUND:\tSolution became worse from:\t{}\tto\t{}".format(optimal_objective_value, new_result))
return optimal_objective_value, optimal_values
if is_result_integer(new_variables):
print("\nFound better integer solution:\tmax clique size:\t{}\n".format(new_result))
return integer_new_result, new_variables
branching_var = branching(new_variables)
# print("BRANCHING:\t\tmax clique float size:\t{}\t\t\t\t\t\t{}".format(new_result, branching_var))
up, down = build_new_constrains(branching_var)
add_constraint(up, model)
optimal_objective_value, optimal_values = branch_and_bound(model,
optimal_objective_value,
optimal_values)
delete_constraint(up['names'][0], model)
add_constraint(down, model)
optimal_objective_value, optimal_values = branch_and_bound(model,
optimal_objective_value,
optimal_values)
delete_constraint(down['names'][0], model)
return optimal_objective_value, optimal_values
def __init_optimization_problem(self):
problem = Cplex()
sense = problem.objective.sense
problem.objective.set_sense(sense=sense.maximize)
variables = self.__build_variables()
# self.__log('Variables: ', variables)
objective = self.__build_objective(variables)
# self.__log('Objective: ', objective)
problem.variables.add(names=variables,
types=['C'] * len(variables),
ub=[1.0] * self.__problem.vertices_num(),
obj=objective)
constraints = self.__build_constraints(variables)
# self.__log('Constraints count: ', len(constraints))
self.__set_constraints(problem, constraints)
self.__optimization_problem = problem
def run_docplex_check_list():
check_platform()
# check requirements
check_import("six")
check_import("enum")
check_import("cloudpickle")
# check cplex
try:
from cplex import Cplex
cpx = Cplex()
del cpx
except ImportError:
print("Cplex DLL not found, if present , you must add it to PYTHNPATH")
# check pandas
try:
import pandas as pd
from pandas import DataFrame, Series
dd = DataFrame({})
except ImportError:
print("-- pandas is not present, some features might be unavailable.")
class SimpleCPLEX:
def __init__(self, cplex=None):
if cplex:
self._model = Cplex(cplex._model)
else:
self._model = Cplex()
self._init_lin()
# to avoid a variable with index 0
self._model.variables.add(
names=['_dummy_'], types=[self._model.variables.type.continuous])
self._var_id = {'_dummy_': 0}
def _init_lin(self):
self._lin = {
'lin_expr': [],
'senses': [],
'rhs': [],
'range_values': [],
'names': []
}
def register_variables(self, prefix, ranges, var_type, lb=None, ub=None):
if not ranges: # None or []
return self._register_variable(prefix, var_type, lb, ub)
variables = {}
for keys in product(*ranges):
name = '_'.join([prefix] + [str(e) for e in keys])
index = self._register_variable(name, var_type, lb, ub)
if len(keys) == 1:
keys = keys[0]
variables[keys] = index
return variables
def _register_variable(self, name, var_type, lb, ub):
self._model.variables.add(names=[name],
types=[var_type],
lb=[] if lb is None else [lb],
ub=[] if ub is None else [ub])
if name in self._var_id:
logger.info('Variable %s is already registered. Overwritten', name)
index = len(self._var_id)
self._var_id[name] = index
return index
def model(self):
return self._model
@property
def parameters(self):
return self._model.parameters
@property
def variables(self):
return self._model.variables
@property
def objective(self):
return self._model.objective
@property
def problem_type(self):
return self._model.problem_type
@property
def solution(self):
return self._model.solution
@property
def version(self):
return self._model.get_version()
def maximize(self):
self._model.objective.set_sense(self._model.objective.sense.maximize)
def minimize(self):
self._model.objective.set_sense(self._model.objective.sense.minimize)
def set_problem_type(self, problem_type):
self._model.set_problem_type(problem_type)
def tune_problem(self, options):
self._model.set_results_stream(None)
self._model.parameters.tune_problem(options)
self._model.set_results_stream(stdout)
def solve(self):
self._model.solve()
def populate(self):
self._model.solve()
self._model.populate_solution_pool()
def variable(self, name):
"""
:param name: variable name
:type name: str
:return: variable index in CPLEX model
:rtype: int
"""
return self._var_id[name]
def get_values(self, lst, idx=None):
if idx:
return self._model.solution.pool.get_values(idx, lst)
else:
return self._model.solution.get_values(lst)
def get_objective_value(self, idx=None):
if idx:
return self._model.solution.pool.get_objective_value(idx)
else:
return self._model.solution.get_objective_value()
@property
def num_solutions(self):
return self._model.solution.pool.get_num()
@staticmethod
def _convert_sense(sense):
# Note: ignore 'R' range case
assert sense in ['E', 'L', 'G', '>=', '=', '==', '<=']
if sense == '<=':
sense = 'L'
elif sense == '=' or sense == '==':
sense = 'E'
elif sense == '>=':
sense = 'G'
return sense
def set_objective(self, lst):
"""
:type lst: list[int or (int, float) or (int, int, float)] or float
"""
if isinstance(lst, float):
self._model.objective.set_offset(lst)
return
linear = []
quad = []
assert isinstance(lst, list)
for e in lst:
assert isinstance(e, int) or isinstance(e, tuple)
if isinstance(e, int):
if e > 0:
linear.append((e, 1))
elif e < 0:
linear.append((-e, -1))
else:
raise RuntimeError('invalid variable ID')
elif len(e) == 2:
linear.append(e)
else:
assert len(e) == 3
e = (min(e[0], e[1]), max(e[0], e[1]), e[2])
quad.append(e)
if linear:
self._model.objective.set_linear(linear)
if quad:
self._model.objective.set_quadratic_coefficients(quad)
@staticmethod
def _convert_coefficients(coef):
"""
Convert 'x', and '-x' into ('x', 1) and ('x', -1), respectively.
:type coef: list[(int, float) or int]
:rtype: (list[int], list[float])
"""
ind = []
val = []
for e in coef:
if isinstance(e, tuple):
assert len(e) == 2
ind.append(e[0])
val.append(e[1])
elif isinstance(e, int):
if e >= 0:
ind.append(e)
val.append(1)
else:
ind.append(-e)
val.append(-1)
else:
raise RuntimeError('unsupported type:' + str(e))
return ind, val
def add_linear_constraint(self, coef, sense, rhs):
"""
:type coef: list[(int, float)]
:type sense: string
:type rhs: float
:rtype: None
"""
if not coef:
logger.warning('empty linear constraint')
return
ind, val = self._convert_coefficients(coef)
sense = self._convert_sense(sense)
c = self._lin
c['lin_expr'].append(SparsePair(ind, val))
c['senses'].append(sense)
c['rhs'].append(rhs)
c['range_values'].append(0)
c['names'].append('c' + str(len(self._lin['names'])))
# logger.debug('%s %s %s %s', c['names'][-1], c['lin_expr'][-1], c['senses'][-1], c['rhs'][-1])
def add_indicator_constraint(self, indvar, complemented, coef, sense, rhs):
"""
:type indvar: int
:type complemented: int
:type coef: list[(int, float)]
:type sense: string
:type rhs: float
:rtype: None
"""
ind, val = self._convert_coefficients(coef)
sense = self._convert_sense(sense)
c = {
'lin_expr': SparsePair(ind, val),
'sense': sense,
'rhs': rhs,
'name': 'i' + str(self._model.indicator_constraints.get_num()),
'indvar': indvar,
'complemented': complemented
}
self._model.indicator_constraints.add(**c)
def add_sos(self, coef):
"""
:type coef: list[(int, float)]
"""
ind, val = self._convert_coefficients(coef)
c = {
'type': '1',
'SOS': SparsePair(ind, val),
'name': 'sos' + str(self._model.SOS.get_num()),
}
self._model.SOS.add(**c)
def add_quadratic_constraint(self, lin, quad, sense, rhs):
"""
:type lin: list[(int, float)]
:type quad: list[(int, int, float)]
:type sense: string
:type rhs: float
:rtype: None
"""
ind, val = self._convert_coefficients(lin)
ind1 = [e[0] for e in quad]
ind2 = [e[1] for e in quad]
val2 = [e[2] for e in quad]
sense = self._convert_sense(sense)
c = {
'lin_expr': SparsePair(ind, val),
'quad_expr': SparseTriple(ind1, ind2, val2),
'sense': sense,
'rhs': rhs,
'name': 'q' + str(self._model.quadratic_constraints.get_num())
}
self._model.quadratic_constraints.add(**c)
def build_model(self):
self._model.linear_constraints.add(**self._lin)
self._init_lin()
def write(self, filename, filetype=''):
self._model.write(filename, filetype)
def create_risk_slim(coef_set, input):
"""
create RiskSLIM MIP object
Parameters
----------
input - dictionary of RiskSLIM parameters and formulation
Returns
-------
mip - RiskSLIM surrogate MIP without 0 cuts
Issues
----
no support for non-integer Lset "values"
only drops intercept index for variable_names that match '(Intercept)'
"""
assert isinstance(coef_set, CoefficientSet)
assert isinstance(input, dict)
# setup printing and loading
function_print_flag = input.get('print_flag', False)
print_from_function = lambda msg: print_log(msg) if function_print_flag else lambda msg: None
update_parameter = lambda pname, pvalue: get_or_set_default(input, pname, pvalue, print_flag = function_print_flag)
# set default parameters
input = update_parameter('w_pos', 1.0)
input = update_parameter('w_neg', 2.0 - input['w_pos'])
input = update_parameter('C_0', 0.01)
input = update_parameter('include_auxillary_variable_for_objval', True)
input = update_parameter('include_auxillary_variable_for_L0_norm', True)
input = update_parameter('loss_min', 0.00)
input = update_parameter('loss_max', float(CPX_INFINITY))
input = update_parameter('L0_min', 0)
input = update_parameter('L0_max', len(coef_set))
input = update_parameter('objval_min', 0.00)
input = update_parameter('objval_max', float(CPX_INFINITY))
input = update_parameter('relax_integer_variables', False)
input = update_parameter('drop_variables', True)
input = update_parameter('tight_formulation', False)
input = update_parameter('set_cplex_cutoffs', True)
# variables
P = len(coef_set)
w_pos, w_neg = input['w_pos'], input['w_neg']
C_0j = np.copy(coef_set.c0)
L0_reg_ind = np.isnan(C_0j)
C_0j[L0_reg_ind] = input['C_0']
C_0j = C_0j.tolist()
C_0_rho = np.copy(C_0j)
trivial_L0_min = 0
trivial_L0_max = np.sum(L0_reg_ind)
rho_ub = list(coef_set.ub)
rho_lb = list(coef_set.lb)
rho_type = ''.join(list(coef_set.vtype))
# calculate min/max values for loss
loss_min = max(0.0, float(input['loss_min']))
loss_max = min(CPX_INFINITY, float(input['loss_max']))
# calculate min/max values for model size
L0_min = max(input['L0_min'], 0.0)
L0_max = min(input['L0_max'], trivial_L0_max)
L0_min = ceil(L0_min)
L0_max = floor(L0_max)
assert L0_min <= L0_max
# calculate min/max values for objval
objval_min = max(input['objval_min'], 0.0)
objval_max = min(input['objval_max'], CPX_INFINITY)
assert objval_min <= objval_max
# include constraint on min/max model size?
nontrivial_L0_min = L0_min > trivial_L0_min
nontrivial_L0_max = L0_max < trivial_L0_max
include_auxillary_variable_for_L0_norm = input['include_auxillary_variable_for_L0_norm'] or \
nontrivial_L0_min or \
nontrivial_L0_max
# include constraint on min/max objective value?
nontrivial_objval_min = objval_min > 0.0
nontrivial_objval_max = objval_max < CPX_INFINITY
include_auxillary_variable_for_objval = input['include_auxillary_variable_for_objval'] or \
nontrivial_objval_min or \
nontrivial_objval_max
has_intercept = '(Intercept)' in coef_set.variable_names
"""
RiskSLIM MIP Formulation
minimize w_pos*loss_pos + w_neg *loss_minus + 0*rho_j + C_0j*alpha_j
such that
L0_min <= L0 <= L0_max
-rho_min * alpha_j < lambda_j < rho_max * alpha_j
L_0 in 0 to P
rho_j in [rho_min_j, rho_max_j]
alpha_j in {0,1}
x = [loss_pos, loss_neg, rho_j, alpha_j]
optional constraints:
objval = w_pos * loss_pos + w_neg * loss_min + sum(C_0j * alpha_j) (required for callback)
L0_norm = sum(alpha_j) (required for callback)
Changes for Tight Formulation (included when input['tight_formulation'] = True):
sigma_j in {0,1} for j s.t. lambda_j has free sign and alpha_j exists
lambda_j >= delta_pos_j if alpha_j = 1 and sigma_j = 1
lambda_j <= -delta_neg_j if alpha_j = 1 and sigma_j = 0
lambda_j >= alpha_j for j such that lambda_j >= 0
lambda_j <= -alpha_j for j such that lambda_j <= 0
"""
# create MIP object
mip = Cplex()
vars = mip.variables
cons = mip.linear_constraints
# set sense
mip.objective.set_sense(mip.objective.sense.minimize)
# add main variables
loss_obj = [w_pos]
loss_ub = [loss_max]
loss_lb = [loss_min]
loss_type = 'C'
loss_names = ['loss']
obj = loss_obj + [0.0] * P + C_0j
ub = loss_ub + rho_ub + [1.0] * P
lb = loss_lb + rho_lb + [0.0] * P
ctype = loss_type + rho_type + 'B' * P
rho_names = ['rho_%d' % j for j in range(P)]
alpha_names = ['alpha_%d' % j for j in range(P)]
varnames = loss_names + rho_names + alpha_names
if include_auxillary_variable_for_objval:
objval_auxillary_name = ['objval']
objval_auxillary_ub = [objval_max]
objval_auxillary_lb = [objval_min]
objval_type = 'C'
print_from_function("adding auxiliary variable for objval s.t. %1.4f <= objval <= %1.4f" % (objval_min, objval_max))
obj += [0.0]
ub += objval_auxillary_ub
lb += objval_auxillary_lb
varnames += objval_auxillary_name
ctype += objval_type
if include_auxillary_variable_for_L0_norm:
L0_norm_auxillary_name = ['L0_norm']
L0_norm_auxillary_ub = [L0_max]
L0_norm_auxillary_lb = [L0_min]
L0_norm_type = 'I'
print_from_function("adding auxiliary variable for L0_norm s.t. %d <= L0_norm <= %d" % (L0_min, L0_max))
obj += [0.0]
ub += L0_norm_auxillary_ub
lb += L0_norm_auxillary_lb
varnames += L0_norm_auxillary_name
ctype += L0_norm_type
if input['relax_integer_variables']:
ctype = ctype.replace('I', 'C')
ctype = ctype.replace('B', 'C')
vars.add(obj = obj, lb = lb, ub = ub, types = ctype, names = varnames)
# 0-Norm LB Constraints:
# lambda_j,lb * alpha_j <= lambda_j <= Inf
# 0 <= lambda_j - lambda_j,lb * alpha_j < Inf
for j in range(P):
cons.add(names = ["L0_norm_lb_" + str(j)],
lin_expr = [SparsePair(ind=[rho_names[j], alpha_names[j]], val=[1.0, -rho_lb[j]])],
senses = "G",
rhs = [0.0])
# 0-Norm UB Constraints:
# lambda_j <= lambda_j,ub * alpha_j
# 0 <= -lambda_j + lambda_j,ub * alpha_j
for j in range(P):
cons.add(names = ["L0_norm_ub_" + str(j)],
lin_expr =[SparsePair(ind=[rho_names[j], alpha_names[j]], val=[-1.0, rho_ub[j]])],
senses = "G",
rhs = [0.0])
# objval_max constraint
# loss_var + sum(C_0j .* alpha_j) <= objval_max
if include_auxillary_variable_for_objval:
print_from_function("adding constraint so that objective value <= " + str(objval_max))
cons.add(names = ["objval_def"],
lin_expr = [SparsePair(ind = objval_auxillary_name + loss_names + alpha_names, val=[-1.0] + loss_obj + C_0j)],
senses = "E",
rhs = [0.0])
# Auxiliary L0_norm variable definition:
# L0_norm = sum(alpha_j)
# L0_norm - sum(alpha_j) = 0
if include_auxillary_variable_for_L0_norm:
cons.add(names = ["L0_norm_def"],
lin_expr = [SparsePair(ind = L0_norm_auxillary_name + alpha_names, val = [1.0] + [-1.0] * P)],
senses = "E",
rhs = [0.0])
# drop L0_norm_lb constraint for any variable with rho_lb >= 0
dropped_variables = []
constraints_to_drop = []
# drop alpha / L0_norm_ub / L0_norm_lb for ('Intercept')
if input['drop_variables']:
# drop L0_norm_ub/lb constraint for any variable with rho_ub/rho_lb >= 0
sign_pos_ind = np.flatnonzero(coef_set.sign > 0)
sign_neg_ind = np.flatnonzero(coef_set.sign < 0)
constraints_to_drop.extend(["L0_norm_lb_" + str(j) for j in sign_pos_ind])
constraints_to_drop.extend(["L0_norm_ub_" + str(j) for j in sign_neg_ind])
# drop alpha for any variable where rho_ub = rho_lb = 0
fixed_value_ind = np.flatnonzero(coef_set.ub == coef_set.lb)
variables_to_drop = ["alpha_" + str(j) for j in fixed_value_ind]
vars.delete(variables_to_drop)
dropped_variables += variables_to_drop
alpha_names = [alpha_names[j] for j in range(P) if alpha_names[j] not in dropped_variables]
if has_intercept:
intercept_idx = coef_set.variable_names.index('(Intercept)')
intercept_alpha_name = 'alpha_' + str(intercept_idx)
vars.delete([intercept_alpha_name])
alpha_names.remove(intercept_alpha_name)
dropped_variables.append(intercept_alpha_name)
print_from_function("dropped L0 indicator for '(Intercept)'")
constraints_to_drop.extend(["L0_norm_ub_" + str(intercept_idx), "L0_norm_lb_" + str(intercept_idx)])
if len(constraints_to_drop) > 0:
constraints_to_drop = list(set(constraints_to_drop))
cons.delete(constraints_to_drop)
# indices
indices = {
'n_variables': vars.get_num(),
'n_constraints': cons.get_num(),
'names': vars.get_names(),
'loss_names': loss_names,
'rho_names': rho_names,
'alpha_names': alpha_names,
'loss': vars.get_indices(loss_names),
'rho': vars.get_indices(rho_names),
'alpha': vars.get_indices(alpha_names),
'L0_reg_ind': L0_reg_ind,
'C_0_rho': C_0_rho,
'C_0_alpha': mip.objective.get_linear(alpha_names) if len(alpha_names) > 0 else [],
}
if include_auxillary_variable_for_objval:
indices.update({
'objval_name': objval_auxillary_name,
'objval': vars.get_indices(objval_auxillary_name)[0],
})
if include_auxillary_variable_for_L0_norm:
indices.update({
'L0_norm_name': L0_norm_auxillary_name,
'L0_norm': vars.get_indices(L0_norm_auxillary_name)[0],
})
# officially change the problem to LP if variables are relaxed
if input['relax_integer_variables']:
old_problem_type = mip.problem_type[mip.get_problem_type()]
mip.set_problem_type(mip.problem_type.LP)
new_problem_type = mip.problem_type[mip.get_problem_type()]
print_from_function("changed problem type from %s to %s" % (old_problem_type, new_problem_type))
if input['set_cplex_cutoffs'] and not input['relax_integer_variables']:
mip.parameters.mip.tolerances.lowercutoff.set(objval_min)
mip.parameters.mip.tolerances.uppercutoff.set(objval_max)
return mip, indices
def Quadratic_constraint(self):
"""Adds Quadratic constraint to the model's Gurobi/Cplex Interface.
(x-mu).T @ inv(cov) @ (x-mu) <= chi-square
Note: This one creates one ellipsoidal constraint for all the metabolites that has non zero or non 'nan' formation energy, irrespective of the magnitude of variance. if the model is infeasible after adding this constraint, refer to util_func.py, find_correlated metabolites to add different ellipsoidal constraints to high variance and normal compounds to avoid possible numerical issues.
Unable to retrieve quadratic constraints in Gurobi model, can see the QC when printed.
:raises NotImplementedError: Implemented only for Gurobi/Cplex interfaces.
:return: [description]
:rtype: [type]
"""
# Pick indices of components present in the current model
model_component_indices = [
i for i in range(self.compound_vector_matrix.shape[1])
if np.any(self.compound_vector_matrix[:, i])
]
# Reduced the compound_vector to contain only the non zero entries
model_compound_vector = self.compound_vector_matrix[:,
model_component_indices]
# Now extract the sub covariance matrix containing only the components present in the model
component_model_covariance = covariance[:, model_component_indices][
model_component_indices, :]
# Now separate the compounds that have variance > 1000 and others to avoid numerical issues
high_variance_indices = np.where(
np.diag(component_model_covariance) > 1000)[0]
low_variance_indices = np.where(
np.diag(component_model_covariance) < 1000)[0]
# Calculate cholesky matrix for two different covariance matrices
if len(low_variance_indices) > 0:
small_component_covariance = component_model_covariance[:, low_variance_indices][
low_variance_indices, :]
cholesky_small_variance = matrix_decomposition(
small_component_covariance)
chi2_value_small = stats.chi2.isf(
q=0.05, df=cholesky_small_variance.shape[1]
) # Chi-square value to map confidence interval
for i in high_variance_indices:
zeros_axis = np.zeros((cholesky_small_variance.shape[1], ))
cholesky_small_variance = np.insert(cholesky_small_variance,
i,
zeros_axis,
axis=0)
metabolite_sphere_small = (
model_compound_vector @ cholesky_small_variance
) # This is a fixed term compound_vector @ cholesky
if len(high_variance_indices) > 0:
large_component_covariance = component_model_covariance[:, high_variance_indices][
high_variance_indices, :] # Covariance matrix for the high variance components
cholesky_large_variance = matrix_decomposition(
large_component_covariance)
chi2_value_high = stats.chi2.isf(
q=0.05, df=cholesky_large_variance.shape[1])
# Insert empty rows for the low_variance_components
for i in low_variance_indices:
zeros_axis = np.zeros((cholesky_large_variance.shape[1], ))
cholesky_large_variance = np.insert(cholesky_large_variance,
i,
zeros_axis,
axis=0)
metabolite_sphere_large = (
model_compound_vector @ cholesky_large_variance
) # This is a fixed term compound_vector @ cholesky
proton_indices = [
self.metabolites.index(metabolite)
for metabolite in self.metabolites
if metabolite.equilibrator_accession is not None
if metabolite.equilibrator_accession.inchi_key == PROTON_INCHI_KEY
] # Get indices of protons in metabolite list to avoid double correcting them for concentrations
if self.solver.__class__.__module__ == "optlang.cplex_interface":
from cplex import Cplex, SparsePair, SparseTriple
# Instantiate Cplex model
cplex_model = Cplex()
rand_str = "".join(
choices(string.ascii_lowercase + string.digits, k=6))
# write cplex model to mps file in random directory and re read
with tempfile.TemporaryDirectory() as td:
temp_filename = os.path.join(td, rand_str + ".mps")
self.solver.problem.write(temp_filename)
cplex_model.read(temp_filename)
# Stop printing output in cplex
cplex_model.set_log_stream(None)
cplex_model.set_error_stream(None)
cplex_model.set_warning_stream(None)
cplex_model.set_results_stream(None)
# Remove the unnecessary variables and constraints
remove_vars = [
var for var in cplex_model.variables.get_names()
if var.startswith("component_") or var.startswith("dG_err_")
] # Remove error variables
remove_constrs = [
cons for cons in cplex_model.linear_constraints.get_names()
if cons.startswith("delG_") or cons.startswith("std_dev_")
] # Remove delG constraint and re-add with component variables
cplex_model.linear_constraints.delete(
remove_constrs) # Removing constr
cplex_model.variables.delete(remove_vars) # Removing Vars
# QC for small variance components
if len(low_variance_indices) > 0:
indices_sphere1 = cplex_model.variables.add(
names=[
"Sphere1_{}".format(i)
for i in range(cholesky_small_variance.shape[1])
],
lb=[-1] * cholesky_small_variance.shape[1],
ub=[1] * cholesky_small_variance.shape[1],
) # Adding independent component variables to the model, store the variable indices
# Add the Sphere constraint
cplex_model.quadratic_constraints.add(
quad_expr=SparseTriple(
ind1=indices_sphere1,
ind2=indices_sphere1,
val=len(indices_sphere1) * [1],
),
sense="L",
rhs=1,
name="unit_normal_small_variance",
)
else:
indices_sphere1 = [
] # Just to adjust the matrix dimensions later
# QC for large variance components
if len(high_variance_indices) > 0:
indices_sphere2 = cplex_model.variables.add(
names=[
"Sphere2_{}".format(i)
for i in range(cholesky_large_variance.shape[1])
],
lb=[-1] * cholesky_large_variance.shape[1],
ub=[1] * cholesky_large_variance.shape[1],
) # Independent large variance components
cplex_model.quadratic_constraints.add(
quad_expr=SparseTriple(
ind1=indices_sphere2,
ind2=indices_sphere2,
val=len(indices_sphere2) * [1],
),
rhs=1,
sense="L",
name="unit_normal_high_variance",
)
else:
indices_sphere2 = [] # Balancing matrix dimensions
concentration_variables = [
"lnc_{}".format(metabolite.id)
for metabolite in self.metabolites
]
# Add the delG constraints
for reaction in self.reactions:
if reaction.id in self.Exclude_reactions:
continue
rxn_stoichiometry = reaction.cal_stoichiometric_matrix()
rxn_stoichiometry = rxn_stoichiometry[np.newaxis, :]
if len(low_variance_indices) > 0:
coefficient_matrix_small_variance = (
np.sqrt(chi2_value_small) *
rxn_stoichiometry @ metabolite_sphere_small
) # Coefficient array for small variance ellipsoid
else:
coefficient_matrix_small_variance = np.array(())
if len(high_variance_indices) > 0:
coefficient_matrix_large_variance = (
np.sqrt(chi2_value_high) *
rxn_stoichiometry @ metabolite_sphere_large
) # Coefficient array for large variance ellipsoid
else:
coefficient_matrix_large_variance = np.array(())
concentration_coefficients = RT * rxn_stoichiometry
concentration_coefficients[0, proton_indices] = 0
coefficients_forward = np.hstack((
np.array((1)),
-1 * concentration_coefficients.flatten(),
-1 * coefficient_matrix_small_variance.flatten(),
-1 * coefficient_matrix_large_variance.flatten(),
))
coefficients_reverse = np.hstack((
np.array((1)),
concentration_coefficients.flatten(),
coefficient_matrix_small_variance.flatten(),
coefficient_matrix_large_variance.flatten(),
))
variable_order_forward = (
["dG_{}".format(reaction.forward_variable.name)] +
concentration_variables + list(indices_sphere1) +
list(indices_sphere2))
variable_order_reverse = (
["dG_{}".format(reaction.reverse_variable.name)] +
concentration_variables + list(indices_sphere1) +
list(indices_sphere2))
rhs = reaction.delG_prime + reaction.delG_transport
cplex_model.linear_constraints.add(
lin_expr=[
SparsePair(
ind=variable_order_forward,
val=coefficients_forward.tolist(),
)
],
senses=["E"],
rhs=[rhs],
names=["delG_{}".format(reaction.forward_variable.name)],
) # delG constraint for forward reaction
cplex_model.linear_constraints.add(
lin_expr=[
SparsePair(
ind=variable_order_reverse,
val=coefficients_reverse.tolist(),
)
],
senses=["E"],
rhs=[-rhs],
names=["delG_{}".format(reaction.reverse_variable.name)],
) # delG constraint for reverse reaction
return cplex_model
elif self.solver.__class__.__module__ == "optlang.gurobi_interface":
from gurobipy import GRB, LinExpr
gurobi_model = self.solver.problem.copy()
# Remove unnecessary variables and constraints and rebuild appropriate ones
remove_vars = [
var for var in gurobi_model.getVars()
if var.VarName.startswith("component_")
or var.VarName.startswith("dG_err_")
]
remove_constrs = [
cons for cons in gurobi_model.getConstrs()
if cons.ConstrName.startswith("delG_")
or cons.ConstrName.startswith("std_dev_")
]
gurobi_model.remove(remove_constrs + remove_vars)
# Add sphere variables for smaller set and larger set separately
if len(low_variance_indices) > 0:
for i in range(cholesky_small_variance.shape[1]):
gurobi_model.addVar(lb=-1,
ub=1,
name="Sphere1_{}".format(i))
gurobi_model.update()
sphere1_variables = [
var for var in gurobi_model.getVars()
if var.VarName.startswith("Sphere1_")
]
gurobi_model.addQConstr(
np.sum(np.square(np.array(sphere1_variables))) <= 1,
name="unit_normal_small_variance",
)
gurobi_model.update()
else:
sphere1_variables = []
# QC for large variance components
if len(high_variance_indices) > 0:
for i in range(cholesky_large_variance.shape[1]):
gurobi_model.addVar(lb=-1,
ub=1,
name="Sphere2_{}".format(i))
gurobi_model.update()
sphere2_variables = [
var for var in gurobi_model.getVars()
if var.VarName.startswith("Sphere2_")
]
gurobi_model.addQConstr(
np.sum(np.square(np.array(sphere2_variables))) <= 1,
name="unit_normal_high_variance",
)
gurobi_model.update()
else:
sphere2_variables = []
# Create a list of metabolite concentration variables
concentration_variables = []
for metabolite in self.metabolites:
varname = "lnc_{}".format(metabolite.id)
conc_var = gurobi_model.getVarByName(varname)
concentration_variables.append(conc_var)
# Add the delG constraints
for reaction in self.reactions:
if reaction.id in self.Exclude_reactions:
continue
rxn_stoichiometry = reaction.cal_stoichiometric_matrix()
rxn_stoichiometry = rxn_stoichiometry[np.newaxis, :]
if len(low_variance_indices) > 0:
coefficient_matrix_small_variance = (
np.sqrt(chi2_value_small) *
rxn_stoichiometry @ metabolite_sphere_small
) # Coefficient array for small variance ellipsoid
else:
coefficient_matrix_small_variance = np.array(())
if len(high_variance_indices) > 0:
coefficient_matrix_large_variance = (
np.sqrt(chi2_value_high) *
rxn_stoichiometry @ metabolite_sphere_large
) # Coefficient array for large variance ellipsoid
else:
coefficient_matrix_large_variance = np.array(())
concentration_coefficients = RT * rxn_stoichiometry
concentration_coefficients[0, proton_indices] = 0
coefficients_forward = np.hstack((
-1 * concentration_coefficients.flatten(),
-1 * coefficient_matrix_small_variance.flatten(),
-1 * coefficient_matrix_large_variance.flatten(),
))
coefficients_reverse = np.hstack((
concentration_coefficients.flatten(),
coefficient_matrix_small_variance.flatten(),
coefficient_matrix_large_variance.flatten(),
))
variable_order = (concentration_variables + sphere1_variables +
sphere2_variables)
delG_err_forward = LinExpr(coefficients_forward.tolist(),
variable_order)
delG_err_reverse = LinExpr(coefficients_reverse.tolist(),
variable_order)
delG_for_var = gurobi_model.getVarByName("dG_{}".format(
reaction.forward_variable.name))
delG_rev_var = gurobi_model.getVarByName("dG_{}".format(
reaction.reverse_variable.name))
rhs = reaction.delG_prime + reaction.delG_transport
gurobi_model.addConstr(
delG_for_var + delG_err_forward,
GRB.EQUAL,
rhs,
name="delG_{}".format(reaction.forward_variable.name),
)
gurobi_model.addConstr(
delG_rev_var + delG_err_reverse,
GRB.EQUAL,
-rhs,
name="delG_{}".format(reaction.reverse_variable.name),
)
gurobi_model.update()
return gurobi_model
else:
raise NotImplementedError("Current solver doesn't support QC")
logging.error(
"Current solver doesnt support problesm of type MIQC")
def __init__(self):
Cplex.__init__(self)
def create_problem(cobra_model, quadratic_component=None, **kwargs):
"""Solver-specific method for constructing a solver problem from
a cobra.Model. This can be tuned for performance using kwargs
"""
# Process parameter defaults
the_parameters = parameter_defaults
if kwargs:
the_parameters = parameter_defaults.copy()
the_parameters.update(kwargs)
if 'relax_b' in the_parameters:
relax_b = the_parameters.pop("relax_b")
warn('need to reimplement relax_b')
relax_b = False
else:
relax_b = False
# Begin problem creation
lp = Cplex()
for k, v in iteritems(the_parameters):
set_parameter(lp, k, v)
objective_coefficients = [
float(x.objective_coefficient) for x in cobra_model.reactions
]
lower_bounds = [_float(x.lower_bound) for x in cobra_model.reactions]
upper_bounds = [_float(x.upper_bound) for x in cobra_model.reactions]
variable_names = cobra_model.reactions.list_attr("id")
variable_kinds = [
variable_kind_dict[x.variable_kind] for x in cobra_model.reactions
]
# Cplex decides that the problem is a MIP if variable_kinds are supplied
# even if there aren't any integers.
if variable_kind_dict['integer'] in variable_kinds:
lp.variables.add(obj=objective_coefficients,
lb=lower_bounds,
ub=upper_bounds,
names=variable_names,
types=variable_kinds)
else:
lp.variables.add(obj=objective_coefficients,
lb=lower_bounds,
ub=upper_bounds,
names=variable_names)
constraint_sense = []
constraint_names = []
constraint_limits = []
for x in cobra_model.metabolites:
constraint_sense.append(x._constraint_sense)
constraint_names.append(x.id)
constraint_limits.append(float(x._bound))
the_linear_expressions = []
# NOTE: This won't work with metabolites that aren't in any reaction
for the_metabolite in cobra_model.metabolites:
variable_list = []
coefficient_list = []
for the_reaction in the_metabolite._reaction:
variable_list.append(the_reaction.id)
coefficient_list.append(
_float(the_reaction._metabolites[the_metabolite]))
the_linear_expressions.append(
SparsePair(ind=variable_list, val=coefficient_list))
# Set objective to quadratic program
if quadratic_component is not None:
set_quadratic_objective(lp, quadratic_component)
if relax_b:
lp.linear_constraints.add(lin_expr=the_linear_expressions,
rhs=constraint_limits,
range_values=list(range_values),
senses=constraint_sense,
names=constraint_names)
else:
lp.linear_constraints.add(lin_expr=the_linear_expressions,
rhs=constraint_limits,
senses=constraint_sense,
names=constraint_names)
# Set the problem type as cplex doesn't appear to do this correctly
problem_type = Cplex.problem_type.LP
if Cplex.variables.type.integer in variable_kinds:
if quadratic_component is not None:
problem_type = Cplex.problem_type.MIQP
else:
problem_type = Cplex.problem_type.MILP
elif quadratic_component is not None:
problem_type = Cplex.problem_type.QP
lp.set_problem_type(problem_type)
return (lp)
class CCOPF:
def __init__(self, conf='conf9', data_dir='data', solver='cplex', verbose=True):
if verbose: print 'loading problem data'
self.conf = conf
self.solver = solver
self.tol = 1e-5
self.verbose = verbose
bus, branch, Bbus, Binv, self.line_nu, self.gen_nu = loadfiles.load(
self.conf, data_dir)
self.data = [bus, branch, Bbus, Binv]
def set_aux(self):
''' Helper variables built from the input data. Call this after
changing the data; don't modify these directly '''
bus, branch, Bbus, Binv = self.data
n = Bbus.shape[0] # number of buses
nl = len(branch) # number of lines
d = [bus[b]['d'] for b in bus] # demand
mu = [bus[b].get('avg', 0) for b in bus] # mean wind power
winds = {b for b in bus if 'std' in bus[b]} # wind source bus IDs
sigma = {b: bus[b]['std'] for b in bus if b in winds} # std wind power
gens = {b for b in bus if bus[b]['type'] == 2} # generator bus IDs
self.aux = [n, nl, d, mu, winds, sigma, gens]
def formulate(self):
self.set_aux()
self.M = Cplex() if self.solver == 'cplex' else Model()
self.addVars()
self.addCons()
self.setObj()
self.p, self.a, self.obj, self.converged, self.tot_its = None,None,None,False,0
if self.verbose:
print 'solving with', self.solver
print '%2s %10s %8s %4s %3s %10s' % ('it', 'worst line', 'viol', 'stat',
'its', 'obj')
def addVars(self):
bus,branch,_,_, n,nl,_,_,_,_,gens = self.data + self.aux
if self.verbose: print 'defining variables'
INF = 1e100
if self.solver == 'cplex':
p = ['p_%d'%i for i in gens]
a = ['a_%d'%i for i in gens]
D = ['D_%d'%i for i in bus]
t = ['t_%d'%i for i in bus]
m = ['m{}'.format(i['id']) for i in branch]
s = ['s{}'.format(i['id']) for i in branch]
self.M.variables.add(names = p + a)
self.M.variables.add(names = D + t, lb = [-INF]*2*n)
#self.M.variables.add(names = m, lb = [-INF]*nl)
#self.M.variables.add(names = s)
self.M.variables.add(names = m + s, lb = [-INF]*2*nl)
D, t = arr(D), arr(t)
self.var = (p, a, D, t, m, s)
else:
p = {i: self.M.addVar(name='pbar_%d'%i) for i in gens}
a = {i: self.M.addVar(name='alpha_%d'%i) for i in gens}
D = {i: self.M.addVar(lb=-INF, name='delta_%d'%i) for i in bus}
t = {i: self.M.addVar(lb=-INF, name='theta_%d'%i) for i in bus}
m = {i['id']: self.M.addVar(lb=-INF, name='fbar{}'.format(i['id'])) for
i in branch}
s = {i['id']: self.M.addVar(lb=-INF, name='std{}'.format(i['id'])) for
i in branch}
self.var = (p, a, D, t, m, s)
self.M.update()
def addCons(self):
bus,branch,Bbus,_, line_nu,gen_nu, n,_,d,mu,winds,sigma,gens = self.data + \
[self.line_nu, self.gen_nu] + self.aux
if self.verbose: print 'defining constraints'
p, a, D, t, m, s = self.var
sumvar = sum(sigma[b]**2 for b in winds)
if self.solver == 'cplex':
ng = len(gens)
self.M.linear_constraints.add(lin_expr = [sp(ind = a, val = [1.0]*ng)],
rhs = [1.0], names=['sum_alpha'])
self.M.linear_constraints.add(lin_expr = [sp(ind = p, val = [1.0]*ng)],
rhs = [sum(d)-sum(mu)], names=['power_balance'])
for i in xrange(n-1):
Bi = Bbus[i,:-1].tocoo()
J,V = Bi.col, Bi.data
if i in gens:
pi, ai, Di, ti = 'p_%d'%i, 'a_%d'%i, 'delta_%d'%i, 'theta_%d'%i
self.M.linear_constraints.add(lin_expr =
[sp(ind = list(D[J])+[ai], val = list(V)+[-1]),
sp(ind = list(t[J])+[pi], val = list(V)+[-1])],
rhs=[0, mu[i]-d[i]], names=[Di, ti], senses='EE')
self.M.linear_constraints.add(lin_expr =
[sp(ind = [pi,ai], val = [1, sumvar**.5*gen_nu]),
sp(ind = [pi,ai], val = [1, -sumvar**.5*gen_nu])],
rhs=[bus[i]['pmax'], bus[i]['pmin']],
names=['gen+_%d'%i, 'gen-_%d'%i], senses='LG')
else:
self.M.linear_constraints.add(lin_expr =
[sp(ind = D[J], val = V), sp(ind = t[J], val = V)],
rhs=[0, mu[i]-d[i]], names=['delta_%d'%i, 'theta_%d'%i],
senses='EE')
self.M.linear_constraints.add(lin_expr =
[sp(ind = [D[n-1]], val = [1.0]), sp(ind = [t[n-1]], val = [1.0])],
rhs = [0,0], names=['delta_%d'%(n-1), 'theta_%d'%(n-1)], senses='EE')
for line in branch:
ID = line['id']
ij = line['ij']
i,j = ij
mij, sij = 'm{}'.format(ID), 's{}'.format(ID)
self.M.linear_constraints.add(lin_expr =
[sp(ind = [mij, t[i], t[j]], val = [1.0, -line['y'], line['y']])],
rhs = [0], names=['line_avg_{}'.format(ID)], senses='E')
self.M.linear_constraints.add(lin_expr =
[sp(ind = [mij, sij], val = [-1.0, line_nu]),
sp(ind = [mij, sij], val = [1.0, line_nu])],
names=['line+_{}'.format(ID), 'line-_{}'.format(ID)],
rhs = [line['fmax'], line['fmax']], senses='LL')
else: # gurobi is much nicer
self.M.addConstr(qs(a[i] for i in gens) == 1, 'sum_alpha_1')
self.M.addConstr(qs(p[i] for i in gens) + sum(mu) == sum(d), 'power_balance')
for i in xrange(n-1):
Bi = Bbus[i,:-1].tocoo()
Bi = zip(Bi.col, Bi.data)
if i in gens:
self.M.addConstr(qs(v*D[j] for j,v in Bi) == a[i], 'delta_%d'%i)
self.M.addConstr(qs(v*t[j] for j,v in Bi) == p[i]+mu[i]-d[i],
'theta_%d'%i)
self.M.addConstr(sumvar**.5*a[i]*gen_nu <= bus[i]['pmax']-p[i],
'gen+_%d'%i)
self.M.addConstr(sumvar**.5*a[i]*gen_nu <= -bus[i]['pmin']+p[i],
'gen-_%d'%i)
else:
self.M.addConstr(qs(v*D[j] for j,v in Bi) == 0, 'delta_%d'%i)
self.M.addConstr(qs(v*t[j] for j,v in Bi) == mu[i]-d[i], 'theta_%d'%i)
self.M.addConstr(D[n-1] == 0, 'delta')
self.M.addConstr(t[n-1] == 0, 'theta')
for line in branch:
ID = line['id']
ij = line['ij']
i,j = ij
self.M.addConstr(m[ID] == line['y']*(t[i]-t[j]),
'line_avg_{}'.format(ID))
self.M.addConstr(s[ID]*line_nu <= line['fmax'] - m[ID],
'line+_{}'.format(ID))
self.M.addConstr(s[ID]*line_nu <= line['fmax'] + m[ID],
'line-_{}'.format(ID))
def setObj(self):
bus,_,_,_, _,_,_,_,winds,sigma,gens = self.data + self.aux
p, a, D, t, m, s = self.var
cost = {b: bus[b]['cost'] for b in bus if b in gens}
sumvar = sum(sigma[b]**2 for b in winds)
if self.solver == 'cplex':
ng = len(gens)
nv = self.M.variables.get_num()
cost = [2*cost[i] for i in gens] + [2*sumvar*cost[i] for i in gens] + \
[0.0]*(nv-2*ng)
self.M.objective.set_quadratic(cost)
self.M.set_results_stream(None)
else:
self.M.setObjective(qs(cost[i]*(p[i]*p[i] + sumvar*a[i]*a[i]) for i in gens))
def solve_subproblem(self):
if self.solver == 'cplex':
self.M.solve()
sol = self.M.solution
self.obj = sol.get_objective_value()
stat = sol.get_status()
if stat == sol.status.optimal:
self.status = 'good'
elif stat == sol.status.num_best:
self.status = 'okay'
else:
self.status = 'fail'
self.statcode = stat
self.its = sol.progress.get_num_barrier_iterations()
self.sol = sol
else:
self.M.params.outputFlag = 0
self.M.optimize()
self.obj = None
if self.M.status == GRB.OPTIMAL:
self.status = 'good'
self.obj = self.M.ObjVal
elif self.M.status == GRB.SUBOPTIMAL:
self.status = 'okay'
self.obj = self.M.ObjVal
else:
self.status = 'fail'
self.statcode = self.M.status
self.its = self.M.BarIterCount
self.tot_its += self.its
def compute_max_viol(self, it):
tol, _,branch,_,Binv, line_nu, n,_,_,_,winds,sigma,gens = [self.tol] + \
self.data + [self.line_nu] + self.aux
p, a, D, t, m, s = self.var
self.worst = {'line': -1, 'val': 0}
for index, line in enumerate(branch):
ID = line['id']
i,j = line['ij']
y, fmax = line['y'], line['fmax']
if self.solver == 'cplex':
Di, Dj, mij = self.sol.get_values([D[i], D[j], 'm{}'.format(ID)])
else:
Di, Dj, mij = D[i].x, D[j].x, m[ID].x
x = arr([y*(Binv[k][i]-Binv[k][j]-Di+Dj)*sigma[k] for k in winds])
violation = (abs(mij)+line_nu*norm(x)-fmax)/fmax
if violation > self.worst['val']:
self.worst = {'line': index, 'val': violation}
if self.worst['val'] < tol:
self.converged = True
self.p, self.a = zeros(n), zeros(n)
for i in gens:
if self.solver == 'cplex':
pi, ai = 'p_%d'%i, 'a_%d'%i
self.p[i], self.a[i] = self.sol.get_values(pi), self.sol.get_values(ai)
else:
self.p[i], self.a[i] = p[i].x, a[i].x
if self.verbose:
i = self.worst['line']
print '%2d %4d->%-4d %8.2e %4s %3d %10.4e' % (it+1, branch[i]['ij'][0],
branch[i]['ij'][1], self.worst['val'], self.status, self.its, self.obj)
def add_cut(self, it):
_,branch,_,Binv, _,_,_,_,winds,sigma,_ = self.data + self.aux
p, a, D, t, m, s = self.var
index = self.worst['line']
ID = branch[index]['id']
i,j = branch[index]['ij']
y = branch[index]['y']
if self.solver == 'cplex':
Di, Dj = self.sol.get_values([D[i], D[j]])
sij = 's{}'.format(ID)
Bij = {k: Binv[k][i]-Binv[k][j] for k in winds}
rhs = y**2*sum([sigma[k]**2*Bij[k]*(Bij[k]-Di+Dj) for k in winds])
coeff = y**2*sum([sigma[k]**2*(Bij[k]-Di+Dj) for k in winds])
x = arr([y*(Bij[k]-Di+Dj)*sigma[k] for k in winds])
#self.M.linear_constraints.add(lin_expr = [sp(ind = [sij], val = [1.0])],
#rhs = [0], names=['cut_help%d'%it], senses='G')
self.M.linear_constraints.add(lin_expr =
[sp(ind = [sij, D[i], D[j]], val = [norm(x), coeff, -coeff])],
rhs = [rhs], names=['cut_%d'%it], senses='G')
else:
x = arr([y*(Binv[k][i]-Binv[k][j]-D[i].x+D[j].x)*sigma[k] for k in winds])
self.M.addConstr(s[ID] >= 0, 'cut_%d_help'%(it+1))
self.M.addConstr(qs(x[l]*y*(Binv[k][i]-Binv[k][j]-D[i]+D[j])*sigma[k]
for l,k in enumerate(winds)) <= s[ID]*norm(x), 'cut_%d'%it)
def set_mean(self, mu):
''' mu is a dictionary {busID: mean wind} '''
bus = self.data[0]
winds = {b for b in bus if 'avg' in bus[b]}
if winds != set(mu.keys()):
print 'error: mu is different'
return None
for i, mean in mu.iteritems():
bus[i]['avg'] = mean
self.set_aux()
def set_std(self, sigma):
''' sigma is a dictionary {busID: std wind} '''
bus = self.data[0]
winds = {b for b in bus if 'std' in bus[b]}
if winds != set(sigma.keys()):
print 'error: sigma is different'
return None
for i, std in sigma.iteritems():
bus[i]['std'] = std
self.set_aux()
def scale_load(self, alpha):
bus = self.data[0]
for i in bus:
bus[i]['d'] *= alpha
self.set_aux()
def scale_lim(self, alpha):
branch = self.data[1]
for line in branch:
line['fmax'] *= alpha
self.set_aux()
#!/usr/bin/env python3
import cplex
from cplex import Cplex
from cplex.exceptions import CplexError
import numpy as np
import matplotlib.pyplot as plt
mip_solver = Cplex()
# mip_solver.set_results_stream(None)
# mip_solver.set_warning_stream(None)
# mip_solver.set_error_stream(None)
#mip_solver.parameters.threads.set(1)
hidden_weights = np.load("hidden_weights.npy")
hidden_bias = np.load("hidden_bias.npy")
output_weights = np.load("output_weights.npy")
output_bias = np.load("output_bias.npy")
input_dim = 28 * 28
hidden_nodes = 20
#mip_solver.objective.set_sense(mip_solver.objective.sense.minimize)
mip_solver.objective.set_sense(mip_solver.objective.sense.maximize)
# Set the value of the output variable as objective function
mip_solver.variables.add(obj=[1],
lb=[-cplex.infinity],
ub=[cplex.infinity],
types="C",
def run_docplex_check_list():
check_platform()
from docplex.version import latest_cplex_major, latest_cplex_minor
cplex_latest_version_as_tuple = (latest_cplex_major, latest_cplex_minor)
diagnostics = []
# check requirements
for rm in ["six", "enum", "cloudpickle"]:
if not check_import(rm):
diagnostics.append(
"Module {0} is missing, run: pip install {0}".format(rm))
# check pandas
try:
import pandas as pd # @UnusedImport
# noinspection PyUnresolvedReferences
from pandas import DataFrame
DataFrame({})
except ImportError:
print("-- pandas is not present, some features might be unavailable.")
from docplex.mp.environment import Environment
Environment().print_information()
# check cplex
try:
# noinspection PyUnresolvedReferences
from cplex import Cplex
cpx = Cplex()
cpxv = cpx.get_version()
cpxvt = tuple(float(x) for x in cpx.get_version().split("."))[:2]
lcpxv = ".".join(str(z) for z in cplex_latest_version_as_tuple)
if cpxvt < cplex_latest_version_as_tuple:
print(
"Warning: Your cplex version {0} is not the latest, {1} is available"
.format(cpxv, lcpxv))
elif cpxvt > cplex_latest_version_as_tuple:
print(
"* Your cplex version {0} is ahead of the latest DOcplex-compatible version {1}, this might not be compatible."
.format(cpxv, lcpxv))
else:
print("* Your cplex version {0} is the latest available".format(
cpxv))
cpx.end()
except ImportError as ie:
Cplex = None
diagnostics.append("No local installation of CPLEX has been found.")
print("Cplex DLL not found, error importing cplex: {0!s}".format(ie))
check_python_path(diagnostics)
# check creation of an empty model...
try:
if Cplex:
# noinspection PyUnresolvedReferences
from docplex.mp.model import Model
Model()
# promotional?
if Model.is_cplex_ce():
print(
"! Cplex promotional version, limited to 1000 variables, 1000 constraints"
)
diagnostics.append(
"Your local CPLEX edition is limited. Consider purchasing a full license."
)
except ImportError:
print("Docplex is not present: cannot import class docplex.mp.model")
diagnostics.append("Your installation of DOcplex may be corrupted.")
except Exception as e:
print(
"Exception raised when creating one model instance: {0!s}".format(
e))
diagnostics.append("Your installation of DOcplex may be corrupted.")
if diagnostics:
print("\n!! diagnostics: {0}".format(len(diagnostics)))
for s in diagnostics:
print(" -- {0}".format(s))
else:
print("> No problem found: you're all set!")
class CplexSolver(Solver):
""" Implements the solver interface using CPLEX. """
def __init__(self, model=None):
Solver.__init__(self)
self.problem = Cplex()
self.status_mapping = {
self.problem.solution.status.optimal:
Status.OPTIMAL,
self.problem.solution.status.optimal_tolerance:
Status.OPTIMAL,
self.problem.solution.status.unbounded:
Status.UNBOUNDED,
self.problem.solution.status.infeasible:
Status.INFEASIBLE,
self.problem.solution.status.infeasible_or_unbounded:
Status.INF_OR_UNB,
self.problem.solution.status.MIP_optimal:
Status.OPTIMAL,
self.problem.solution.status.MIP_unbounded:
Status.UNBOUNDED,
self.problem.solution.status.MIP_infeasible:
Status.INFEASIBLE,
self.problem.solution.status.MIP_infeasible_or_unbounded:
Status.INF_OR_UNB
}
self.vartype_mapping = {
VarType.BINARY: self.problem.variables.type.binary,
VarType.INTEGER: self.problem.variables.type.integer,
VarType.CONTINUOUS: self.problem.variables.type.continuous
}
self.parameter_mapping = {
Parameter.TIME_LIMIT: self.problem.parameters.timelimit,
Parameter.FEASIBILITY_TOL:
self.problem.parameters.simplex.tolerances.feasibility,
Parameter.OPTIMALITY_TOL:
self.problem.parameters.simplex.tolerances.optimality,
Parameter.INT_FEASIBILITY_TOL:
self.problem.parameters.mip.tolerances.integrality,
Parameter.MIP_ABS_GAP:
self.problem.parameters.mip.tolerances.mipgap,
Parameter.MIP_REL_GAP:
self.problem.parameters.mip.tolerances.absmipgap,
Parameter.POOL_SIZE: self.problem.parameters.mip.limits.populate,
Parameter.POOL_GAP: self.problem.parameters.mip.pool.relgap
}
self.set_logging()
self.set_parameters(default_parameters)
# self.problem.parameters.randomseed.set(0)
# self.problem.parameters.mip.strategy.search.set(1)
self._cached_lin_obj = {}
self._cached_sense = None
self._cached_lower_bounds = {}
self._cached_upper_bounds = {}
self._cached_vars = []
self._cached_constrs = []
if model:
self.build_problem(model)
def add_variable(self,
var_id,
lb=None,
ub=None,
vartype=VarType.CONTINUOUS,
persistent=True,
update_problem=True):
""" Add a variable to the current problem.
Arguments:
var_id (str): variable identifier
lb (float): lower bound
ub (float): upper bound
vartype (VarType): variable type (default: CONTINUOUS)
persistent (bool): if the variable should be reused for multiple calls (default: true)
update_problem (bool): update problem immediately
"""
if update_problem:
self.add_variables([var_id], [lb], [ub], [vartype])
else:
self._cached_vars.append((var_id, lb, ub, vartype))
if not persistent:
self.temp_vars.add(var_id)
def add_variables(self, var_ids, lbs, ubs, vartypes):
""" Add multiple variables to the current problem.
Arguments:
var_ids (list): variable identifier
lbs (list): lower bounds
ubs (list): upper bounds
vartypes (list): variable types (default: CONTINUOUS)
"""
lbs = [lb if lb is not None else -infinity for lb in lbs]
ubs = [ub if ub is not None else infinity for ub in ubs]
if set(vartypes) == {VarType.CONTINUOUS}:
self.problem.variables.add(names=var_ids, lb=lbs, ub=ubs)
else:
vartypes = [self.vartype_mapping[vartype] for vartype in vartypes]
self.problem.variables.add(names=var_ids,
lb=lbs,
ub=ubs,
types=vartypes)
self.var_ids.extend(var_ids)
self._cached_lower_bounds.update(dict(zip(var_ids, lbs)))
self._cached_upper_bounds.update(dict(zip(var_ids, ubs)))
self._cached_lin_obj.update({var_id: 0.0 for var_id in var_ids})
def add_constraint(self,
constr_id,
lhs,
sense='=',
rhs=0,
persistent=True,
update_problem=True):
""" Add a constraint to the current problem.
Arguments:
constr_id (str): constraint identifier
lhs (dict): variables and respective coefficients
sense (str): constraint sense (any of: '<', '=', '>'; default '=')
rhs (float): right-hand side of equation (default: 0)
persistent (bool): if the variable should be reused for multiple calls (default: True)
update_problem (bool): update problem immediately
"""
if update_problem:
self.add_constraints([constr_id], [lhs], [sense], [rhs])
else:
self._cached_constrs.append((constr_id, lhs, sense, rhs))
if not persistent:
self.temp_constrs.add(constr_id)
def add_constraints(self, constr_ids, lhs, senses, rhs):
""" Add a list of constraints to the current problem.
Arguments:
constr_ids (list): constraint identifiers
lhs (list): variables and respective coefficients
senses (list): constraint senses (default: '=')
rhs (list): right-hand side of equations (default: 0)
"""
map_sense = {'=': 'E', '<': 'L', '>': 'G'}
exprs = [
SparsePair(ind=list(constr.keys()), val=list(constr.values()))
for constr in lhs
]
senses = [map_sense[sense] for sense in senses]
self.problem.linear_constraints.add(lin_expr=exprs,
senses=senses,
rhs=rhs,
names=constr_ids)
self.constr_ids.extend(constr_ids)
def remove_variable(self, var_id):
""" Remove a variable from the current problem.
Arguments:
var_id (str): variable identifier
"""
self.remove_variables([var_id])
def remove_variables(self, var_ids):
""" Remove variables from the current problem.
Arguments:
var_ids (list): variable identifiers
"""
found = []
for var_id in var_ids:
if var_id in self.var_ids:
found.append(var_id)
self.var_ids.remove(var_id)
self.problem.variables.delete(found)
def remove_constraint(self, constr_id):
""" Remove a constraint from the current problem.
Arguments:
constr_id (str): constraint identifier
"""
self.remove_constraints([constr_id])
def remove_constraints(self, constr_ids):
""" Remove constraints from the current problem.
Arguments:
constr_ids (list): constraint identifiers
"""
found = []
for constr_id in constr_ids:
if constr_id in self.constr_ids:
found.append(constr_id)
self.constr_ids.remove(constr_id)
self.problem.linear_constraints.delete(found)
def update(self):
""" Update internal structure. Used for efficient lazy updating. """
if self._cached_vars:
var_ids, lbs, ubs, vartypes = list(zip(*self._cached_vars))
self.add_variables(var_ids, lbs, ubs, vartypes)
self._cached_vars = []
if self._cached_constrs:
constr_ids, lhs, senses, rhs = list(zip(*self._cached_constrs))
self.add_constraints(constr_ids, lhs, senses, rhs)
self._cached_constrs = []
def set_objective(self, linear=None, quadratic=None, minimize=True):
""" Set a predefined objective for this problem.
Args:
linear (dict): linear coefficients (optional)
quadratic (dict): quadratic coefficients (optional)
minimize (bool): solve a minimization problem (default: True)
Notes:
Setting the objective is optional. It can also be passed directly when calling **solve**.
"""
if linear:
updated_coeffs = {}
for var_id in self.var_ids:
if var_id in linear and linear[var_id] != self._cached_lin_obj[
var_id]:
updated_coeffs[var_id] = linear[var_id]
if var_id not in linear and self._cached_lin_obj[var_id] != 0.0:
updated_coeffs[var_id] = 0.0
if updated_coeffs:
self.problem.objective.set_linear(list(updated_coeffs.items()))
self._cached_lin_obj.update(updated_coeffs)
if quadratic:
self.problem.objective.set_quadratic(
[0.0] * len(self.var_ids)) #TODO: is this really necessary ?
quad_coeffs = [(r_id1, r_id2, coeff)
for (r_id1, r_id2), coeff in quadratic.items()]
self.problem.objective.set_quadratic_coefficients(quad_coeffs)
if minimize != self._cached_sense:
if minimize:
sense = self.problem.objective.sense.minimize
else:
sense = self.problem.objective.sense.maximize
self.problem.objective.set_sense(sense)
self._cached_sense = minimize
def build_problem(self, model):
""" Create problem structure for a given model.
Arguments:
model : CBModel
"""
var_ids = list(model.reactions.keys())
lbs = [rxn.lb for rxn in model.reactions.values()]
ubs = [rxn.ub for rxn in model.reactions.values()]
var_types = [VarType.CONTINUOUS] * len(var_ids)
self.add_variables(var_ids, lbs, ubs, var_types)
constr_ids = list(model.metabolites.keys())
table = model.metabolite_reaction_lookup(force_recalculate=True)
lhs = list(table.values())
senses = ['='] * len(constr_ids)
rhs = [0] * len(constr_ids)
self.add_constraints(constr_ids, lhs, senses, rhs)
def solve(self,
linear=None,
quadratic=None,
minimize=None,
model=None,
constraints=None,
get_values=True,
get_shadow_prices=False,
get_reduced_costs=False,
pool_size=0,
pool_gap=None):
""" Solve the optimization problem.
Arguments:
linear (dict): linear objective (optional)
quadratic (dict): quadratic objective (optional)
minimize (bool): solve a minimization problem (default: True)
model (CBModel): model (optional, leave blank to reuse previous model structure)
constraints (dict): additional constraints (optional)
get_values (bool or list): set to false for speedup if you only care about the objective value (default: True)
get_shadow_prices (bool): return shadow prices if available (default: False)
get_reduced_costs (bool): return reduced costs if available (default: False)
pool_size (int): calculate solution pool of given size (only for MILP problems)
pool_gap (float): maximum relative gap for solutions in pool (optional)
Returns:
Solution: solution
"""
if model:
self.build_problem(model)
problem = self.problem
if constraints:
changed_lb, changed_ub = self.temporary_bounds(constraints)
self.set_objective(linear, quadratic, minimize)
# from datetime import datetime
# self.write_to_file(f"{datetime.now()}.lp")
#run the optimization
if pool_size == 0:
problem.solve()
status = self.status_mapping.get(problem.solution.get_status(),
Status.UNKNOWN)
message = str(problem.solution.get_status_string())
if status == Status.OPTIMAL:
fobj = problem.solution.get_objective_value()
values, shadow_prices, reduced_costs = None, None, None
if get_values:
if isinstance(get_values, Iterable):
get_values = list(get_values)
values = OrderedDict(
zip(get_values,
problem.solution.get_values(get_values)))
else:
values = OrderedDict(
zip(self.var_ids, problem.solution.get_values()))
if get_shadow_prices:
shadow_prices = OrderedDict(
zip(self.constr_ids,
problem.solution.get_dual_values(self.constr_ids)))
if get_reduced_costs:
reduced_costs = OrderedDict(
zip(self.var_ids,
problem.solution.get_reduced_costs(self.var_ids)))
solution = Solution(status, message, fobj, values,
shadow_prices, reduced_costs)
else:
solution = Solution(status, message)
else:
pool_pmap = {
'SolnPoolIntensity': problem.parameters.mip.pool.intensity,
'PopulateLim': problem.parameters.mip.limits.populate,
'SolnPoolCapacity': problem.parameters.mip.pool.capacity,
'SolnPoolReplace': problem.parameters.mip.pool.replace,
'SolnPoolGap': problem.parameters.mip.pool.relgap,
'SolnPoolAGap': problem.parameters.mip.pool.absgap
}
default_params = {
'SolnPoolIntensity': 3,
'PopulateLim': 10 * pool_size,
'SolnPoolCapacity': pool_size,
'SolnPoolReplace': 1
}
for param, val in default_params.items():
pool_pmap[param].set(val)
if pool_gap:
pool_pmap['SolnPoolGap'].set(pool_gap)
problem.populate_solution_pool()
status = self.status_mapping.get(problem.solution.get_status(),
Status.UNKNOWN)
if status == Status.OPTIMAL or status == Status.UNKNOWN:
solution = self.get_solution_pool(get_values)
else:
solution = []
if constraints:
self.reset_bounds(changed_lb, changed_ub)
return solution
def temporary_bounds(self, constraints):
lower_bounds, upper_bounds = {}, {}
lb_new, ub_new = {}, {}
def _dict_diff(dict1, dict2):
return set(dict1.items()) - set(dict2.items())
for r_id, x in constraints.items():
if r_id in self.var_ids:
lb, ub = x if isinstance(x, tuple) else (x, x)
lower_bounds[r_id] = lb if lb is not None else -infinity
upper_bounds[r_id] = ub if ub is not None else infinity
else:
warnings.warn(
"Constrained variable '{}' not previously declared".format(
r_id), RuntimeWarning)
if lower_bounds != self._cached_lower_bounds:
lb_new = _dict_diff(lower_bounds, self._cached_lower_bounds)
if len(lb_new) > 0:
self.problem.variables.set_lower_bounds(lb_new)
if upper_bounds != self._cached_upper_bounds:
ub_new = _dict_diff(upper_bounds, self._cached_upper_bounds)
if len(ub_new) > 0:
self.problem.variables.set_upper_bounds(ub_new)
return lb_new, ub_new
def get_solution_pool(self, get_values=True):
""" Return a solution pool for MILP problems.
Must be called after using solve with pool_size argument > 0.
Arguments:
get_values (bool or list): set to false for speedup if you only care about the objective value (default: True)
Returns:
list: list of Solution objects
"""
pool = self.problem.solution.pool
solutions = []
for i in range(pool.get_num()):
obj = pool.get_objective_value(i)
# TODO: remove all OrderedDicts when migrating to python 3.7
# values = OrderedDict([(r_id, pool.get_values(i, r_id)) for r_id in self.var_ids])
if get_values:
if isinstance(get_values, Iterable):
get_values = list(get_values)
values = dict(
zip(get_values, pool.get_values(i, get_values)))
else:
values = dict(zip(self.var_ids, pool.get_values(i)))
else:
values = None
sol = Solution(fobj=obj, values=values)
solutions.append(sol)
return solutions
def reset_bounds(self, updated_lb, updated_ub):
if updated_lb:
lb_old = [(r_id, self._cached_lower_bounds[r_id])
for r_id, _ in updated_lb]
self.problem.variables.set_lower_bounds(lb_old)
if updated_ub:
ub_old = [(r_id, self._cached_upper_bounds[r_id])
for r_id, _ in updated_ub]
self.problem.variables.set_upper_bounds(ub_old)
def set_lower_bounds(self, bounds_dict):
self.problem.variables.set_lower_bounds([
(var_id, lb if lb is not None else -infinity)
for var_id, lb in bounds_dict.items()
])
def set_upper_bounds(self, bounds_dict):
self.problem.variables.set_lower_bounds([
(var_id, ub if ub is not None else infinity)
for var_id, ub in bounds_dict.items()
])
def set_bounds(self, bounds_dict):
self.problem.variables.set_lower_bounds([
(var_id, bounds[0] if bounds[0] is not None else -infinity)
for var_id, bounds in bounds_dict.items()
])
self.problem.variables.set_upper_bounds([
(var_id, bounds[1] if bounds[1] is not None else infinity)
for var_id, bounds in bounds_dict.items()
])
def update_coefficient(self, coeff, var_id, value):
self.problem.linear_constraints.set_coefficients([(coeff, var_id,
value)])
def update_coefficients(self, coefficients):
self.problem.linear_constraints.set_coefficients(coefficients)
def set_parameter(self, parameter, value):
""" Set a parameter value for this optimization problem
Arguments:
parameter (Parameter): parameter type
value (float): parameter value
"""
if parameter in self.parameter_mapping:
self.parameter_mapping[parameter].set(value)
else:
raise Exception('Parameter unknown (or not yet supported).')
def set_logging(self, enabled=False):
""" Enable or disable log output:
Arguments:
enabled (bool): turn logging on (default: False)
"""
if enabled:
self.problem.set_log_stream(sys.stdout)
self.problem.set_error_stream(sys.stderr)
self.problem.set_warning_stream(sys.stderr)
self.problem.set_results_stream(sys.stdout)
else:
self.problem.set_log_stream(None)
self.problem.set_error_stream(None)
self.problem.set_warning_stream(None)
self.problem.set_results_stream(None)
def write_to_file(self, filename):
""" Write problem to file:
Arguments:
filename (str): file path
"""
self.problem.write(filename)
def variability_legacy_cplex(
model,
variable_list=None,
params=False,
):
"""Custom function to perform TVA on MIQC problem using gurobi.
Parameters
----------
model : multitfa.core.tmodel
multitfa model after thermodynamic constraints are added
variable_list : List, optional
List of variables to perform TVA on, by default None
params : Bool, optional
If True sets the Timelimit option to 300 sec and reduced the mip gap to 0.005
Returns
-------
pd.DataFrame
Dataframe of min max ranges of variables
Raises
------
ValueError
[description]
"""
# Instead of copying the whole model, just copy the cplex solver object by writing to a file and reading again.
from cplex import Cplex, SparsePair
tmp_dir = (os.path.normpath(os.path.dirname(os.path.abspath(__file__))) +
os.sep + os.pardir + os.sep + "tmp")
if not os.path.exists(tmp_dir):
os.makedirs(tmp_dir)
# Instantiate Cplex model
cplex_model = Cplex()
rand_str = "".join(choices(string.ascii_lowercase + string.digits, k=6))
# write cplex model to mps file and re read
with tempfile.TemporaryDirectory() as td:
temp_filename = os.path.join(td, rand_str + ".mps")
model.cplex_interface.write(temp_filename)
cplex_model.read(temp_filename)
cplex_model.set_log_stream(None)
cplex_model.set_error_stream(None)
cplex_model.set_warning_stream(None)
cplex_model.set_results_stream(None)
if params:
# print("lol")
cplex_model.parameters.mip.tolerances.mipgap = 0.005
cplex_model.parameters.timelimit = 300
cplex_model.parameters.mip.limits.probetime = 300
# Make shorts for sense
max_sense = cplex_model.objective.sense.maximize
min_sense = cplex_model.objective.sense.minimize
if variable_list == None:
variables = model.cplex_interface.variables.get_names()
else:
variables = [var for var in variable_list]
vars_list_cplex = cplex_model.variables.get_names()
fluxes_min = np.empty(len(variables))
fluxes_max = np.empty(len(variables))
rxn_name = list()
rxn_ids = [rxn.id for rxn in model.reactions]
for i in range(len(variables)):
# Reset objective vector for each iteration
for varname in vars_list_cplex:
cplex_model.objective.set_linear(varname, 0)
# if the variable is reactions optimize for forward - reverse variables else optimize for the variable
if variables[i] in rxn_ids:
rxn = model.reactions.get_by_id(variables[i])
cplex_model.objective.set_linear([(rxn.forward_variable.name, 1),
(rxn.reverse_variable.name, -1)])
else:
cplex_model.objective.set_linear(variables[i], 1)
rxn_name.append(variables[i])
# minimization
cplex_model.objective.set_sense(min_sense)
cplex_model.solve()
objective_value = cplex_model.solution.get_objective_value()
fluxes_min[i] = objective_value
# maximiztion
cplex_model.objective.set_sense(max_sense)
cplex_model.solve()
objective_value = cplex_model.solution.get_objective_value()
fluxes_max[i] = objective_value
return DataFrame({
"minimum": Series(index=rxn_name, data=fluxes_min),
"maximum": Series(index=rxn_name, data=fluxes_max),
})
#!/usr/bin/env python3
import cplex
from cplex import Cplex
from cplex.exceptions import CplexError
import numpy as np
import matplotlib.pyplot as plt
mip_solver = Cplex()
mip_solver.set_results_stream(None)
mip_solver.set_warning_stream(None)
mip_solver.set_error_stream(None)
# mip_solver.parameters.threads.set(1)
hidden_weights = [np.load("hidden_weights_1.npy"), np.load("hidden_weights_2.npy")]
hidden_bias = [np.load("hidden_bias_1.npy"), np.load("hidden_bias_2.npy")]
output_weights = np.load("output_weights.npy")
output_bias = np.load("output_bias.npy")
#mip_solver.objective.set_sense(mip_solver.objective.sense.minimize)
input_dim = 28*28
hidden_nodes = [10,10]
output_nodes = 10
# output variables
mip_solver.variables.add(
lb = [-cplex.infinity]*output_nodes,
ub = [cplex.infinity]*output_nodes,
types = "C"*output_nodes,
def _optimize_cplex(cobra_model,
new_objective=None,
objective_sense='maximize',
min_norm=0,
the_problem=None,
tolerance_optimality=1e-6,
tolerance_feasibility=1e-6,
tolerance_integer=1e-9,
tolerance_barrier=1e-8,
error_reporting=None,
print_solver_time=False,
lp_method=1,
lp_parallel=0,
copy_problem=False,
relax_b=None,
quadratic_component=None,
reuse_basis=True,
update_problem_reaction_bounds=True):
"""Uses the ILOG/CPLEX (www.ibm.com/software/integration/optimization/cplex-optimizer/)
optimizer to perform an optimization on cobra_model for the objective_coefficients in
cobra_model._objective_coefficients based on the objective sense.
cobra_model: A cobra.Model object
new_objective: Reaction, String, or Integer referring to a reaction in
cobra_model.reactions to set as the objective. Currently, only supports single
objective coeffients. Will expand to include mixed objectives.
objective_sense: 'maximize' or 'minimize'
min_norm: not implemented
the_problem: None or a problem object for the specific solver that can be used to hot
start the next solution.
tolerance_optimality: Solver tolerance for optimality.
tolerance_feasibility: Solver tolerance for feasibility.
error_reporting: None or True to disable or enable printing errors encountered
when trying to find the optimal solution.
print_solver_time: False or True. Indicates if the time to calculate the solution
should be displayed.
quadratic_component: None or
scipy.sparse.dok of dim(len(cobra_model.reactions),len(cobra_model.reactions))
If not None:
Solves quadratic programming problems for cobra_models of the form:
minimize: 0.5 * x' * quadratic_component * x + cobra_model._objective_coefficients' * x
such that,
cobra_model._lower_bounds <= x <= cobra_model._upper_bounds
cobra_model._S * x (cobra_model._constraint_sense) cobra_model._b
reuse_basis: Boolean. If True and the_problem is a model object for the solver,
attempt to hot start the solution.
update_problem_reaction_bounds: Boolean. Set to True if you're providing the_problem
and you've modified reaction bounds on your cobra_model since creating the_problem. Only
necessary for CPLEX
method for linear optimization: 0 = automatic
1 = primal simplex, 2 = dual simplex, 3 = network simplex,
4 = barrier, 5 = sifting, 6 = concurrent dual, barrier, and primal
lp.solve() with Salmonella model:
cold start: 0.05 seconds
hot start: 0.05 seconds (slow due to copying the LP)
"""
if relax_b is not None:
raise Exception('Need to reimplement constraint relaxation')
from numpy import array, nan, zeros
from cobra.flux_analysis.objective import update_objective
from cobra.solvers.legacy import status_dict, variable_kind_dict
if error_reporting == 'time' or print_solver_time:
from time import time
start_time = time()
try:
from cplex import Cplex, SparsePair
variable_kind_dict = eval(variable_kind_dict['cplex'])
status_dict = eval(status_dict['cplex'])
except ImportError as e:
import sys
if 'wrong architecture' in e[0] and sys.maxsize > 2**32:
print 'CPLEX python API is not 64-bit. please contact your IBM representative'
else:
print e
if new_objective and new_objective != 'update problem':
update_objective(cobra_model, new_objective)
if the_problem == None or the_problem in ['return', 'setup', 'parallel'] \
or not isinstance(the_problem, Cplex):
lp = Cplex()
#Using the new objects
#NOTE: This might be slow
objective_coefficients = []
lower_bounds = []
upper_bounds = []
variable_names = []
variable_kinds = []
[(objective_coefficients.append(x.objective_coefficient),
lower_bounds.append(x.lower_bound),
upper_bounds.append(x.upper_bound), variable_names.append(x.id),
variable_kinds.append(variable_kind_dict[x.variable_kind]))
for x in cobra_model.reactions]
#Cplex decides that the problem is a MIP if variable_kinds are supplied
#even if there aren't any integers.
if Cplex.variables.type.integer in variable_kinds:
lp.variables.add(obj=objective_coefficients,
lb=lower_bounds,
ub=upper_bounds,
names=variable_names,
types=variable_kinds)
else:
lp.variables.add(obj=objective_coefficients,
lb=lower_bounds,
ub=upper_bounds,
names=variable_names)
if relax_b:
raise Exception('need to reimplement relax_b')
## range_values = zeros(len(cobra_model.metabolites))
## b_values = array([x._bound for x in cobra_model.metabolties])
## for the_nonzero in list(b_values.nonzero()[0]):
## range_values[the_nonzero] = -relax_b
constraint_sense = []
constraint_names = []
constraint_limits = []
[(constraint_sense.append(x._constraint_sense),
constraint_names.append(x.id), constraint_limits.append(x._bound))
for x in cobra_model.metabolites]
the_linear_expressions = []
#NOTE: This won't work with metabolites that aren't in any reaction
for the_metabolite in cobra_model.metabolites:
variable_list = []
coefficient_list = []
for the_reaction in the_metabolite._reaction:
variable_list.append(the_reaction.id)
coefficient_list.append(
the_reaction._metabolites[the_metabolite])
the_linear_expressions.append(
SparsePair(ind=variable_list, val=coefficient_list))
if quadratic_component is not None:
if not hasattr(quadratic_component, 'todok'):
raise Exception(
'quadratic component must be a scipy.sparse type array')
quadratic_component_scaled = quadratic_component.todok()
lp.parameters.emphasis.numerical.set(1)
for k, v in quadratic_component_scaled.items():
lp.objective.set_quadratic_coefficients(
int(k[0]), int(k[1]), v)
if relax_b:
lp.linear_constraints.add(lin_expr=the_linear_expressions,
rhs=constraint_limits,
range_values=list(range_values),
senses=constraint_sense,
names=constraint_names)
else:
lp.linear_constraints.add(lin_expr=the_linear_expressions,
rhs=constraint_limits,
senses=constraint_sense,
names=constraint_names)
if error_reporting == 'time':
print 'setup new problem: ' + repr(time() - start_time)
start_time = time()
#Set the problem type as cplex doesn't appear to do this correctly
problem_type = Cplex.problem_type.LP
if Cplex.variables.type.integer in variable_kinds:
if quadratic_component is not None:
problem_type = Cplex.problem_type.MIQP
else:
problem_type = Cplex.problem_type.MILP
elif quadratic_component is not None:
problem_type = Cplex.problem_type.QP
lp.set_problem_type(problem_type)
else:
if copy_problem:
lp = Cplex(the_problem)
if error_reporting == 'time':
print 'copy problem: ' + repr(time() - start_time)
start_time = time()
else:
lp = the_problem
if new_objective:
lp.objective.set_linear([(x.id, float(x.objective_coefficient))
for x in cobra_model.reactions])
if error_reporting == 'time':
print 'set lp objective: ' + repr(time() - start_time)
start_time = time()
#SPEED THIS UP
if update_problem_reaction_bounds:
lp.variables.set_upper_bounds([(x.id, float(x.upper_bound))
for x in cobra_model.reactions])
lp.variables.set_lower_bounds([(x.id, float(x.lower_bound))
for x in cobra_model.reactions])
if error_reporting == 'time':
print 'changed all bounds: ' + repr(time() - start_time)
start_time = time()
if objective_sense == 'maximize':
lp.objective.set_sense(lp.objective.sense.maximize)
else:
lp.objective.set_sense(lp.objective.sense.minimize)
if tolerance_optimality < 1e-10:
lp.parameters.simplex.perturbation.constant.set(1)
lp.parameters.simplex.pgradient.set(1)
lp.parameters.emphasis.memory.set(1)
#lp.parameters.simplex.tolerances.markowitz.set(.01)
lp.parameters.advance.set(2)
lp.parameters.simplex.tolerances.optimality.set(tolerance_optimality)
lp.parameters.simplex.tolerances.feasibility.set(tolerance_feasibility)
if lp.get_problem_type() in [
Cplex.problem_type.LP, Cplex.problem_type.MILP
]:
lp.parameters.lpmethod.set(lp_method)
elif lp.get_problem_type() in [
Cplex.problem_type.QP, Cplex.problem_type.MIQP
]:
lp.parameters.qpmethod.set(lp_method)
if lp_parallel > 1:
lp.parameters.threads.set(lp_parallel)
#lp.parameters.parallel.set(lp_parallel)
lp.parameters.barrier.convergetol.set(tolerance_barrier)
if the_problem == 'setup':
return lp
if not error_reporting:
lp.set_results_stream(None)
lp.set_warning_stream(None)
if print_solver_time:
start_time = time()
if not isinstance(the_problem, Cplex):
#TODO: set tolerance
lp.solve()
# Solve this LP with the simplex method. Takes about 0.2 s without hot start
lp.status = lp.solution.status[lp.solution.get_status()]
if lp.status in status_dict:
status = status_dict[lp.status]
else:
status = 'failed'
else:
if isinstance(the_problem, Cplex) and reuse_basis:
try:
the_basis = the_problem.solution.basis.get_basis()
lp.start.set_basis(the_basis[0], the_basis[1])
#TODO: Determine whether the primal or dual works best for the
#problem of interest. For the ME matrix the primal appears to
#work best
lp_method = 1
lp.parameters.preprocessing.presolve.set(0)
lp.parameters.lpmethod.set(lp_method)
except:
print 'no basis in the_problem'
#TODO: set tolerance and time limit
#lp.parameters.timelimit.set()
lp.solve()
#If the solver takes more than 0.1 s with a hot start it is likely stuck
lp.status = lp.solution.status[lp.solution.get_status()]
if lp.status in status_dict:
status = status_dict[lp.status]
else:
status = 'failed'
if status != 'optimal':
#Cycle through the different solver options, if a solution is not found
for lp_method in (1, 2, 3, 4, 5, 6):
lp = optimize_cplex(
cobra_model,
new_objective=new_objective,
objective_sense=objective_sense,
min_norm=min_norm,
the_problem=None,
print_solver_time=print_solver_time,
tolerance_optimality=tolerance_optimality,
tolerance_feasibility=tolerance_feasibility,
lp_method=lp_method,
quadratic_component=quadratic_component)['the_problem']
lp.status = lp.solution.status[lp.solution.get_status()]
if lp.status in status_dict:
status = status_dict[lp.status]
else:
status = 'failed'
if status == 'optimal':
break
if error_reporting == 'time':
print 'solver time: ' + repr(
time() - start_time) + ' with method ' + repr(lp_method)
start_time = time()
if print_solver_time:
print 'cplex time: %f' % (time() - start_time)
#TODO: It might be able to speed this up a little.
if status == 'optimal':
objective_value = lp.solution.get_objective_value()
#This can be sped up a little
x_dict = dict(zip(lp.variables.get_names(), lp.solution.get_values()))
x = array(lp.solution.get_values())
x = x.reshape(x.shape[0], 1)
#MIP's don't have duals
if lp.get_problem_type() in (Cplex.problem_type.MIQP,
Cplex.problem_type.MILP):
y = y_dict = None
else:
y_dict = dict(
zip(lp.linear_constraints.get_names(),
lp.solution.get_dual_values()))
y = array(lp.solution.get_dual_values())
y = y.reshape(y.shape[0], 1)
else:
x = y = x_dict = y_dict = objective_value = None
if error_reporting:
print 'cplex failed: %s' % lp.status
cobra_model.solution = the_solution = Solution(objective_value,
x=x,
x_dict=x_dict,
status=status,
y=y,
y_dict=y_dict)
solution = {'the_problem': lp, 'the_solution': the_solution}
return solution
def _optimize_cplex(cobra_model, new_objective=None, objective_sense='maximize',
min_norm=0, the_problem=None,
tolerance_optimality=1e-6, tolerance_feasibility=1e-6, tolerance_integer=1e-9,
tolerance_barrier=1e-8,error_reporting=None,
print_solver_time=False, lp_method=1, lp_parallel=0, copy_problem=False,
relax_b=None, quadratic_component=None, reuse_basis=True,
update_problem_reaction_bounds=True):
"""Uses the ILOG/CPLEX (www.ibm.com/software/integration/optimization/cplex-optimizer/)
optimizer to perform an optimization on cobra_model for the objective_coefficients in
cobra_model._objective_coefficients based on the objective sense.
cobra_model: A cobra.Model object
new_objective: Reaction, String, or Integer referring to a reaction in
cobra_model.reactions to set as the objective. Currently, only supports single
objective coeffients. Will expand to include mixed objectives.
objective_sense: 'maximize' or 'minimize'
min_norm: not implemented
the_problem: None or a problem object for the specific solver that can be used to hot
start the next solution.
tolerance_optimality: Solver tolerance for optimality.
tolerance_feasibility: Solver tolerance for feasibility.
error_reporting: None or True to disable or enable printing errors encountered
when trying to find the optimal solution.
print_solver_time: False or True. Indicates if the time to calculate the solution
should be displayed.
quadratic_component: None or
scipy.sparse.dok of dim(len(cobra_model.reactions),len(cobra_model.reactions))
If not None:
Solves quadratic programming problems for cobra_models of the form:
minimize: 0.5 * x' * quadratic_component * x + cobra_model._objective_coefficients' * x
such that,
cobra_model._lower_bounds <= x <= cobra_model._upper_bounds
cobra_model._S * x (cobra_model._constraint_sense) cobra_model._b
reuse_basis: Boolean. If True and the_problem is a model object for the solver,
attempt to hot start the solution.
update_problem_reaction_bounds: Boolean. Set to True if you're providing the_problem
and you've modified reaction bounds on your cobra_model since creating the_problem. Only
necessary for CPLEX
method for linear optimization: 0 = automatic
1 = primal simplex, 2 = dual simplex, 3 = network simplex,
4 = barrier, 5 = sifting, 6 = concurrent dual, barrier, and primal
lp.solve() with Salmonella model:
cold start: 0.05 seconds
hot start: 0.05 seconds (slow due to copying the LP)
"""
if relax_b is not None:
raise Exception('Need to reimplement constraint relaxation')
from numpy import array, nan, zeros
from cobra.flux_analysis.objective import update_objective
from cobra.solvers.legacy import status_dict, variable_kind_dict
if error_reporting == 'time' or print_solver_time:
from time import time
start_time = time()
try:
from cplex import Cplex, SparsePair
variable_kind_dict = eval(variable_kind_dict['cplex'])
status_dict = eval(status_dict['cplex'])
except ImportError as e:
import sys
if 'wrong architecture' in e[0] and sys.maxsize > 2**32:
print 'CPLEX python API is not 64-bit. please contact your IBM representative'
else:
print e
if new_objective and new_objective != 'update problem':
update_objective(cobra_model, new_objective)
if the_problem == None or the_problem in ['return', 'setup', 'parallel'] \
or not isinstance(the_problem, Cplex):
lp = Cplex()
#Using the new objects
#NOTE: This might be slow
objective_coefficients = []
lower_bounds = []
upper_bounds = []
variable_names = []
variable_kinds = []
[(objective_coefficients.append(x.objective_coefficient),
lower_bounds.append(x.lower_bound),
upper_bounds.append(x.upper_bound),
variable_names.append(x.id),
variable_kinds.append(variable_kind_dict[x.variable_kind]))
for x in cobra_model.reactions]
#Cplex decides that the problem is a MIP if variable_kinds are supplied
#even if there aren't any integers.
if Cplex.variables.type.integer in variable_kinds:
lp.variables.add(obj=objective_coefficients,
lb=lower_bounds,
ub=upper_bounds,
names=variable_names,
types=variable_kinds)
else:
lp.variables.add(obj=objective_coefficients,
lb=lower_bounds,
ub=upper_bounds,
names=variable_names)
if relax_b:
raise Exception('need to reimplement relax_b')
## range_values = zeros(len(cobra_model.metabolites))
## b_values = array([x._bound for x in cobra_model.metabolties])
## for the_nonzero in list(b_values.nonzero()[0]):
## range_values[the_nonzero] = -relax_b
constraint_sense = []
constraint_names = []
constraint_limits = []
[(constraint_sense.append(x._constraint_sense),
constraint_names.append(x.id),
constraint_limits.append(x._bound))
for x in cobra_model.metabolites]
the_linear_expressions = []
#NOTE: This won't work with metabolites that aren't in any reaction
for the_metabolite in cobra_model.metabolites:
variable_list = []
coefficient_list = []
for the_reaction in the_metabolite._reaction:
variable_list.append(the_reaction.id)
coefficient_list.append(the_reaction._metabolites[the_metabolite])
the_linear_expressions.append(SparsePair(ind=variable_list,
val=coefficient_list))
if quadratic_component is not None:
if not hasattr(quadratic_component, 'todok'):
raise Exception('quadratic component must be a scipy.sparse type array')
quadratic_component_scaled = quadratic_component.todok()
lp.parameters.emphasis.numerical.set(1)
for k, v in quadratic_component_scaled.items():
lp.objective.set_quadratic_coefficients(int(k[0]), int(k[1]), v)
if relax_b:
lp.linear_constraints.add(lin_expr=the_linear_expressions,
rhs=constraint_limits,
range_values=list(range_values),
senses=constraint_sense,
names=constraint_names)
else:
lp.linear_constraints.add(lin_expr=the_linear_expressions,
rhs=constraint_limits,
senses=constraint_sense,
names=constraint_names)
if error_reporting == 'time':
print 'setup new problem: ' + repr(time()-start_time)
start_time = time()
#Set the problem type as cplex doesn't appear to do this correctly
problem_type = Cplex.problem_type.LP
if Cplex.variables.type.integer in variable_kinds:
if quadratic_component is not None:
problem_type = Cplex.problem_type.MIQP
else:
problem_type = Cplex.problem_type.MILP
elif quadratic_component is not None:
problem_type = Cplex.problem_type.QP
lp.set_problem_type(problem_type)
else:
if copy_problem:
lp = Cplex(the_problem)
if error_reporting == 'time':
print 'copy problem: ' + repr(time()-start_time)
start_time = time()
else:
lp = the_problem
if new_objective:
lp.objective.set_linear([(x.id, float(x.objective_coefficient))
for x in cobra_model.reactions])
if error_reporting == 'time':
print 'set lp objective: ' + repr(time()-start_time)
start_time = time()
#SPEED THIS UP
if update_problem_reaction_bounds:
lp.variables.set_upper_bounds([(x.id, float(x.upper_bound))
for x in cobra_model.reactions])
lp.variables.set_lower_bounds([(x.id, float(x.lower_bound))
for x in cobra_model.reactions])
if error_reporting == 'time':
print 'changed all bounds: ' + repr(time()-start_time)
start_time = time()
if objective_sense == 'maximize':
lp.objective.set_sense(lp.objective.sense.maximize)
else:
lp.objective.set_sense(lp.objective.sense.minimize)
if tolerance_optimality < 1e-10:
lp.parameters.simplex.perturbation.constant.set(1)
lp.parameters.simplex.pgradient.set(1)
lp.parameters.emphasis.memory.set(1)
#lp.parameters.simplex.tolerances.markowitz.set(.01)
lp.parameters.advance.set(2)
lp.parameters.simplex.tolerances.optimality.set(tolerance_optimality)
lp.parameters.simplex.tolerances.feasibility.set(tolerance_feasibility)
if lp.get_problem_type() in [Cplex.problem_type.LP,
Cplex.problem_type.MILP]:
lp.parameters.lpmethod.set(lp_method)
elif lp.get_problem_type() in [Cplex.problem_type.QP,
Cplex.problem_type.MIQP]:
lp.parameters.qpmethod.set(lp_method)
if lp_parallel > 1:
lp.parameters.threads.set(lp_parallel)
#lp.parameters.parallel.set(lp_parallel)
lp.parameters.barrier.convergetol.set(tolerance_barrier)
if the_problem == 'setup':
return lp
if not error_reporting:
lp.set_results_stream(None)
lp.set_warning_stream(None)
if print_solver_time:
start_time = time()
if not isinstance(the_problem, Cplex):
#TODO: set tolerance
lp.solve()
# Solve this LP with the simplex method. Takes about 0.2 s without hot start
lp.status = lp.solution.status[lp.solution.get_status()]
if lp.status in status_dict:
status = status_dict[lp.status]
else:
status = 'failed'
else:
if isinstance(the_problem, Cplex) and reuse_basis:
try:
the_basis = the_problem.solution.basis.get_basis()
lp.start.set_basis(the_basis[0],the_basis[1])
#TODO: Determine whether the primal or dual works best for the
#problem of interest. For the ME matrix the primal appears to
#work best
lp_method = 1
lp.parameters.preprocessing.presolve.set(0)
lp.parameters.lpmethod.set(lp_method)
except:
print 'no basis in the_problem'
#TODO: set tolerance and time limit
#lp.parameters.timelimit.set()
lp.solve()
#If the solver takes more than 0.1 s with a hot start it is likely stuck
lp.status = lp.solution.status[lp.solution.get_status()]
if lp.status in status_dict:
status = status_dict[lp.status]
else:
status = 'failed'
if status != 'optimal':
#Cycle through the different solver options, if a solution is not found
for lp_method in (1, 2, 3, 4, 5, 6):
lp = optimize_cplex(cobra_model, new_objective=new_objective,
objective_sense=objective_sense,
min_norm=min_norm, the_problem=None,
print_solver_time=print_solver_time,
tolerance_optimality=tolerance_optimality,
tolerance_feasibility=tolerance_feasibility,
lp_method=lp_method,
quadratic_component=quadratic_component)['the_problem']
lp.status = lp.solution.status[lp.solution.get_status()]
if lp.status in status_dict:
status = status_dict[lp.status]
else:
status = 'failed'
if status == 'optimal':
break
if error_reporting == 'time':
print 'solver time: ' + repr(time()-start_time) + ' with method ' + repr(lp_method)
start_time = time()
if print_solver_time:
print 'cplex time: %f'%(time() - start_time)
x = []
x_dict = {}
#TODO: It might be able to speed this up a little.
if status == 'optimal':
objective_value = lp.solution.get_objective_value()
#This can be sped up a little
x_dict = dict(zip(lp.variables.get_names(),
lp.solution.get_values()))
x = array(lp.solution.get_values())
x = x.reshape(x.shape[0],1)
#MIP's don't have duals
if lp.get_problem_type() in (Cplex.problem_type.MIQP,
Cplex.problem_type.MILP):
y = y_dict = None
else:
y_dict = dict(zip(lp.linear_constraints.get_names(),
lp.solution.get_dual_values()))
y = array(lp.solution.get_dual_values())
y = y.reshape(y.shape[0],1)
else:
x = y = x_dict = y_dict = objective_value = None
if error_reporting:
print 'cplex failed: %s'%lp.status
the_solution = Solution(objective_value, x=x, x_dict=x_dict,
status=status, y=y, y_dict=y_dict)
solution = {'the_problem': lp, 'the_solution': the_solution}
return solution
def qps_cplex(H, c, A, l, u, xmin, xmax, x0, opt):
"""Quadratic Program Solver based on CPLEX.
A wrapper function providing a PYPOWER standardized interface for using
C{cplexqp} or C{cplexlp} to solve the following QP (quadratic programming)
problem::
min 1/2 X'*H*x + c'*x
x
subject to::
l <= A*x <= u (linear constraints)
xmin <= x <= xmax (variable bounds)
Inputs (all optional except C{H}, C{c}, C{A} and C{l}):
- C{H} : matrix (possibly sparse) of quadratic cost coefficients
- C{c} : vector of linear cost coefficients
- C{A, l, u} : define the optional linear constraints. Default
values for the elements of L and U are -Inf and Inf, respectively.
- C{xmin, xmax} : optional lower and upper bounds on the
C{x} variables, defaults are -Inf and Inf, respectively.
- C{x0} : optional starting value of optimization vector C{x}
- C{opt} : optional options structure with the following fields,
all of which are also optional (default values shown in parentheses)
- C{verbose} (0) - controls level of progress output displayed
- 0 = no progress output
- 1 = some progress output
- 2 = verbose progress output
- C{cplex_opt} - options dict for CPLEX, value in
verbose overrides these options
- C{problem} : The inputs can alternatively be supplied in a single
C{problem} dict with fields corresponding to the input arguments
described above: C{H, c, A, l, u, xmin, xmax, x0, opt}
Outputs:
- C{x} : solution vector
- C{f} : final objective function value
- C{exitflag} : CPLEXQP/CPLEXLP exit flag
(see C{cplexqp} and C{cplexlp} documentation for details)
- C{output} : CPLEXQP/CPLEXLP output dict
(see C{cplexqp} and C{cplexlp} documentation for details)
- C{lmbda} : dict containing the Langrange and Kuhn-Tucker
multipliers on the constraints, with fields:
- mu_l - lower (left-hand) limit on linear constraints
- mu_u - upper (right-hand) limit on linear constraints
- lower - lower bound on optimization variables
- upper - upper bound on optimization variables
@author: Ray Zimmerman (PSERC Cornell)
"""
##----- input argument handling -----
## gather inputs
if isinstance(H, dict): ## problem struct
p = H
if 'opt' in p: opt = p['opt']
if 'x0' in p: x0 = p['x0']
if 'xmax' in p: xmax = p['xmax']
if 'xmin' in p: xmin = p['xmin']
if 'u' in p: u = p['u']
if 'l' in p: l = p['l']
if 'A' in p: A = p['A']
if 'c' in p: c = p['c']
if 'H' in p: H = p['H']
else: ## individual args
assert H is not None
assert c is not None
assert A is not None
assert l is not None
if opt is None:
opt = {}
# if x0 is None:
# x0 = array([])
# if xmax is None:
# xmax = array([])
# if xmin is None:
# xmin = array([])
## define nx, set default values for missing optional inputs
if len(H) == 0 or not any(any(H)):
if len(A) == 0 and len(xmin) == 0 and len(xmax) == 0:
stderr.write(
'qps_cplex: LP problem must include constraints or variable bounds\n'
)
else:
if len(A) > 0:
nx = shape(A)[1]
elif len(xmin) > 0:
nx = len(xmin)
else: # if len(xmax) > 0
nx = len(xmax)
else:
nx = shape(H)[0]
if len(c) == 0:
c = zeros(nx)
if len(A) > 0 and (len(l) == 0 or all(l == -Inf)) and \
(len(u) == 0 or all(u == Inf)):
A = None ## no limits => no linear constraints
nA = shape(A)[0] ## number of original linear constraints
if len(u) == 0: ## By default, linear inequalities are ...
u = Inf * ones(nA) ## ... unbounded above and ...
if len(l) == 0:
l = -Inf * ones(nA) ## ... unbounded below.
if len(xmin) == 0: ## By default, optimization variables are ...
xmin = -Inf * ones(nx) ## ... unbounded below and ...
if len(xmax) == 0:
xmax = Inf * ones(nx) ## ... unbounded above.
if len(x0) == 0:
x0 = zeros(nx)
## default options
if 'verbose' in opt:
verbose = opt['verbose']
else:
verbose = 0
#if 'max_it' in opt:
# max_it = opt['max_it']
#else:
# max_it = 0
## split up linear constraints
ieq = find(abs(u - l) <= EPS) ## equality
igt = find(u >= 1e10 & l > -1e10) ## greater than, unbounded above
ilt = find(l <= -1e10 & u < 1e10) ## less than, unbounded below
ibx = find((abs(u - l) > EPS) & (u < 1e10) & (l > -1e10))
Ae = A[ieq, :]
be = u[ieq]
Ai = r_[A[ilt, :], -A[igt, :], A[ibx, :] - A[ibx, :]]
bi = r_[u[ilt], -l[igt], u[ibx], -l[ibx]]
## grab some dimensions
nlt = len(ilt) ## number of upper bounded linear inequalities
ngt = len(igt) ## number of lower bounded linear inequalities
nbx = len(ibx) ## number of doubly bounded linear inequalities
## set up options struct for CPLEX
if 'cplex_opt' in opt:
cplex_opt = cplex_options(opt['cplex_opt'])
else:
cplex_opt = cplex_options
cplex = Cplex('null')
vstr = cplex.getVersion
s, e, tE, m, t = re.compile(vstr, '(\d+\.\d+)\.')
vnum = int(t[0][0])
vrb = max([0, verbose - 1])
cplex_opt['barrier']['display'] = vrb
cplex_opt['conflict']['display'] = vrb
cplex_opt['mip']['display'] = vrb
cplex_opt['sifting']['display'] = vrb
cplex_opt['simplex']['display'] = vrb
cplex_opt['tune']['display'] = vrb
if vrb and (vnum > 12.2):
cplex_opt['diagnostics'] = 'on'
#if max_it:
# cplex_opt. ## not sure what to set here
if len(Ai) == 0 and len(Ae) == 0:
unconstrained = 1
Ae = sparse((1, nx))
be = 0
else:
unconstrained = 0
## call the solver
if verbose:
methods = [
'default', 'primal simplex', 'dual simplex', 'network simplex',
'barrier', 'sifting', 'concurrent'
]
if len(H) == 0 or not any(any(H)):
if verbose:
stdout.write('CPLEX Version %s -- %s LP solver\n' %
(vstr, methods[cplex_opt['lpmethod'] + 1]))
x, f, eflag, output, lam = \
cplexlp(c, Ai, bi, Ae, be, xmin, xmax, x0, cplex_opt)
else:
if verbose:
stdout.write('CPLEX Version %s -- %s QP solver\n' %
(vstr, methods[cplex_opt['qpmethod'] + 1]))
## ensure H is numerically symmetric
if H != H.T:
H = (H + H.T) / 2
x, f, eflag, output, lam = \
cplexqp(H, c, Ai, bi, Ae, be, xmin, xmax, x0, cplex_opt)
## check for empty results (in case optimization failed)
if len(x) == 0:
x = NaN * zeros(nx)
if len(f) == 0:
f = NaN
if len(lam) == 0:
lam['ineqlin'] = NaN * zeros(len(bi))
lam['eqlin'] = NaN * zeros(len(be))
lam['lower'] = NaN * zeros(nx)
lam['upper'] = NaN * zeros(nx)
mu_l = NaN * zeros(nA)
mu_u = NaN * zeros(nA)
else:
mu_l = zeros(nA)
mu_u = zeros(nA)
if unconstrained:
lam['eqlin'] = array([])
## negate prices depending on version
if vnum < 12.3:
lam['eqlin'] = -lam['eqlin']
lam['ineqlin'] = -lam['ineqlin']
## repackage lambdas
kl = find(lam.eqlin < 0) ## lower bound binding
ku = find(lam.eqlin > 0) ## upper bound binding
mu_l[ieq[kl]] = -lam['eqlin'][kl]
mu_l[igt] = lam['ineqlin'][nlt + arange(ngt)]
mu_l[ibx] = lam['ineqlin'][nlt + ngt + nbx + arange(nbx)]
mu_u[ieq[ku]] = lam['eqlin'][ku]
mu_u[ilt] = lam['ineqlin'][:nlt]
mu_u[ibx] = lam['ineqlin'][nlt + ngt + arange(nbx)]
lmbda = {
'mu_l': mu_l,
'mu_u': mu_u,
'lower': lam.lower,
'upper': lam.upper
}
return x, f, eflag, output, lmbda
def create_problem(cobra_model, quadratic_component=None, **kwargs):
"""Solver-specific method for constructing a solver problem from
a cobra.Model. This can be tuned for performance using kwargs
"""
# Process parameter defaults
the_parameters = parameter_defaults
if kwargs:
the_parameters = parameter_defaults.copy()
the_parameters.update(kwargs)
if 'relax_b' in the_parameters:
relax_b = the_parameters.pop("relax_b")
warn('need to reimplement relax_b')
relax_b = False
else:
relax_b = False
# Begin problem creation
lp = Cplex()
for k, v in iteritems(the_parameters):
set_parameter(lp, k, v)
objective_coefficients = [float(x.objective_coefficient)
for x in cobra_model.reactions]
lower_bounds = [float(x.lower_bound) for x in cobra_model.reactions]
upper_bounds = [float(x.upper_bound) for x in cobra_model.reactions]
variable_names = cobra_model.reactions.list_attr("id")
variable_kinds = [variable_kind_dict[x.variable_kind] for x
in cobra_model.reactions]
# Cplex decides that the problem is a MIP if variable_kinds are supplied
# even if there aren't any integers.
if variable_kind_dict['integer'] in variable_kinds:
lp.variables.add(obj=objective_coefficients,
lb=lower_bounds,
ub=upper_bounds,
names=variable_names,
types=variable_kinds)
else:
lp.variables.add(obj=objective_coefficients,
lb=lower_bounds,
ub=upper_bounds,
names=variable_names)
## if relax_b:
## range_values = zeros(len(cobra_model.metabolites))
## b_values = array([x._bound for x in cobra_model.metabolties])
## for the_nonzero in list(b_values.nonzero()[0]):
## range_values[the_nonzero] = -relax_b
constraint_sense = []
constraint_names = []
constraint_limits = []
[(constraint_sense.append(x._constraint_sense),
constraint_names.append(x.id),
constraint_limits.append(float(x._bound)))
for x in cobra_model.metabolites]
the_linear_expressions = []
#NOTE: This won't work with metabolites that aren't in any reaction
for the_metabolite in cobra_model.metabolites:
variable_list = []
coefficient_list = []
for the_reaction in the_metabolite._reaction:
variable_list.append(the_reaction.id)
coefficient_list.append(float(the_reaction._metabolites[the_metabolite]))
the_linear_expressions.append(SparsePair(ind=variable_list,
val=coefficient_list))
# Set objective to quadratic program
if quadratic_component is not None:
set_quadratic_objective(lp, quadratic_component)
if relax_b:
lp.linear_constraints.add(lin_expr=the_linear_expressions,
rhs=constraint_limits,
range_values=list(range_values),
senses=constraint_sense,
names=constraint_names)
else:
lp.linear_constraints.add(lin_expr=the_linear_expressions,
rhs=constraint_limits,
senses=constraint_sense,
names=constraint_names)
#Set the problem type as cplex doesn't appear to do this correctly
problem_type = Cplex.problem_type.LP
if Cplex.variables.type.integer in variable_kinds:
if quadratic_component is not None:
problem_type = Cplex.problem_type.MIQP
else:
problem_type = Cplex.problem_type.MILP
elif quadratic_component is not None:
problem_type = Cplex.problem_type.QP
lp.set_problem_type(problem_type)
return(lp)
from cplex import Cplex, infinity
from cplex.exceptions import CplexError
problem = Cplex()
problem.objective.set_sense(problem.objective.sense.minimize)
objective = [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0]
constraints_matrix = [
[1.0, 1.0, 1.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[1.0, 1.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
[1.0, 0.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.0, 0.0, 0.0, 0.0],
[0.0, 0.0, 1.0, 1.0, 1.0, 1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0],
[1.0, 1.0, 1.0, 1.0, 1.0, 0.0, 0.0, 0.0, 0.0, 1.0, 1.0, 0.0],
[0.0, 0.0, 1.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 0.0, 1.0, 0.0],
[0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0, 1.0, 0.0, 0.0, 0.0, 1.0],
[0.0, 0.0, 1.0, 0.0, 0.0, 1.0, 1.0, 1.0, 1.0, 0.0, 1.0, 1.0],
[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 1.0, 1.0, 1.0, 1.0],
[0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0, 1.0, 1.0, 0.0],
[0.0, 0.0, 0.0, 1.0, 1.0, 1.0, 0.0, 1.0, 1.0, 1.0, 1.0, 0.0],
[0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 1.0, 1.0, 0.0, 0.0, 1.0],
]
col_names = [
"x1", "x2", "x3", "x4", "x5", "x6", "x7", "x8", "x9", "x10", "x11", "x12"
]
row_names = [
"c1", "c2", "c3", "c4", "c5", "c6", "c7", "c8", "c9", "c10", "c11", "c12"
]
def fair_partial_assignment_lp_solver(df, centers, color_flag, alpha, beta,
cost_fun_string):
# There are primarily five steps:
# 1. Initiate a model for cplex
# 2. Declare if it is minimization or maximization problem
# 3. Add variables to the model. The variables are generally named.
# The upper bounds and lower bounds on the range for the variables
# are also mentioned at this stage. The coefficient of the objective
# functions are also entered at this step
# 4. Add the constraints to the model. The constraint matrix, denoted by A,
# can be added in three ways - row wise, column wise or non-zero entry wise.
# 5. Finally, call the solver.
# Step 1. Initiate a model for cplex.
print("Initializing Cplex model")
problem = Cplex()
# Step 2. Declare that this is a minimization problem
problem.objective.set_sense(problem.objective.sense.minimize)
# Step 3. Declare and add variables to the model. The function
# prepare_to_add_variables (points, center) prepares all the
# required information for this stage.
#
# objective: a list of coefficients (float) in the linear objective function
# lower bounds: a list of floats containing the lower bounds for each variable
# upper bounds: a list of floats containing the upper bounds for each variable
# variable_name: a list of strings that contains the name of the variables
print("Starting to add variables...")
print("HERE???")
t1 = time.monotonic()
objective, lower_bounds, upper_bounds, variable_names = prepare_to_add_variables(
df, centers, cost_fun_string)
problem.variables.add(obj=objective,
lb=lower_bounds,
ub=upper_bounds,
names=variable_names)
t2 = time.monotonic()
print("Completed. Time for creating and adding variable = {}".format(t2 -
t1))
# Step 4. Declare and add constraints to the model.
# There are few ways of adding constraints: rwo wise, col wise and non-zero entry wise.
# Assume the constraint matrix is A. We add the constraints row wise.
# The function prepare_to_add_constraints_by_entry(points,center,colors,alpha,beta)
# prepares the required data for this step.
#
# constraints_row: Encoding of each row of the constraint matrix
# senses: a list of strings that identifies whether the corresponding constraint is
# an equality or inequality. "E" : equals to (=), "L" : less than (<=), "G" : greater than equals (>=)
# rhs: a list of floats corresponding to the rhs of the constraints.
# constraint_names: a list of string corresponding to the name of the constraint
print("Starting to add constraints...")
t1 = time.monotonic()
objects_returned = prepare_to_add_constraints(df, centers, color_flag,
beta, alpha)
constraints_row, senses, rhs, constraint_names = objects_returned
problem.linear_constraints.add(lin_expr=constraints_row,
senses=senses,
rhs=rhs,
names=constraint_names)
t2 = time.monotonic()
print(
"Completed. Time for creating and adding constraints = {}".format(t2 -
t1))
# Optional: We can set various parameters to optimize the performance of the lp solver
# As an example, the following sets barrier method as the lp solving method
# The other available methods are: auto, primal, dual, sifting, concurrent
#problem.parameters.lpmethod.set(problem.parameters.lpmethod.values.barrier)
return problem, objective
def solveit(max_duration, ads):
# get the winners
winners, losers = select_winners(ads, max_duration)
# add a fake winner if needed.
total_dur = sum([winner.duration for winner in winners])
if total_dur < max_duration:
winners.add(Ad(bid = 0.0001, duration =max_duration - total_dur))
# get the upper contour for the losers
bar = [i[1] for i in get_coutour(losers, max_duration, True)]
print len(winners)
winners = sorted(winners, key = lambda winner: (winner.duration, winner.bid))
id2num = {id(winner):k for k, winner in enumerate(winners)}
prob = Cplex()
lb = [bar[winner.duration] for winner in winners]
ub = [winner.bid for winner in winners]
prob.objective.set_sense(prob.objective.sense.minimize)
obj = [1] * len(winners)
prob.variables.add(obj, lb, ub)
# monotone requirement (optional)
# for i, x in enumerate(winners):
# ceil = 1e9
# for k, y in enumerate(winners[i+1:]):
# j = i + 1 + k
# if y.bid == x.bid:
# if x.duration == y.duration:
# prob.linear_constraints.add([[[i, j], [1,-1]]], ['E'], [0])
# else:
# prob.linear_constraints.add([[[i, j],[1,-1]]],['L'],[0])
# break
# if y.bid > x.bid and y.bid < ceil:
# ceil = y.bid
# prob.linear_constraints.add([[[i,j],[1,-1]]],['L'],[0])
n_iteration = 0
last_obj = -1
while True:
n_iteration += 1
prob.solve()
obj = prob.solution.get_objective_value()
last_obj = obj
prices = prob.solution.get_values()
for i, winner in enumerate(winners):
winner.price = prices[i]
lower_coutour = get_coutour(winners, max_duration, False)
finish = True
for k, tmp in enumerate(lower_coutour):
ads, min_value = tmp
if len(ads) == 0:
continue
if min_value < bar[k]:
tmp = [id2num[id(ad)] for ad in ads]
print prob.linear_constraints.get_num()
prob.linear_constraints.add([[[id2num[id(ad)] for ad in ads], [1] * len(ads)]],\
['G'], [bar[k]])
finish = False
if finish:
break
return sum(prob.solution.get_values()), len(winners), n_iteration
