def deterministic_solve_function(model_data, output_directory, mipgap=.01):
import Core.Solvers.MSSP.Deterministic_Solver as Solve
MSSP_Solution = Solve.deterministic_PRDP_solve_with_return(
mipgap, model_data, output_directory)
return MSSP_Solution
def KDA_Solve_Function(model_data,
output_directory,
fixed_parameters={},
_opts=[],
solver='cplex',
mipgap=.001):
## Run KDA to get lower bound, this returns the result and the ENPV location
import Core.Solvers.KDA.KDA_Solution_Class as Solve
import Core.Solvers.KDA.Evaluate_KDA_PRDP as Evaluate_KDA
Solution = Solve.KDA(model_data, solver, mipgap, output_directory, _opts,
fixed_parameters)
#### Calculate the equivalent ENPV
results = Solution.output['results']
sp_realizations = Solution.output['sub_problem_realizations']
Evaluated_Solution = Evaluate_KDA.KDA_PRDP_results(model_data, results,
sp_realizations,
output_directory)
### Write results to Consolidated file
import Core.output_write as output_write
output_write._write(Evaluated_Solution, Solution, output_directory, 'kda',
'')
import Core.Bounding_Class as Bounding_Class
return_value = Bounding_Class.Bkda(Evaluated_Solution, Solution)
return return_value
def solve_function(solve_options):
### Define Empty Options Variable
optcmds = {}
### Split list of Parsed Commands on equal signs
for i in solve_options[1:]:
try:
cmd = i.split('=',1)
optcmds[cmd[0]] = cmd[1]
except:
try:
optcmds['none'].append(i)
except:
optcmds['none'] = [i]
### Lists of acceptable inputs
list_of_solvers = ['kda','mtssp','mssp','saa','eoss']
_cmds = ['linear-relaxation','min_step','max_step','all_time']
solver_options = ['min_solve','max_solve', 'probabilistic', 'penalty', 'SAA', 'PH', 'deterministic_initializations']
MSSP_opts = []
_opts = []
### Sort the options chosen
for i in optcmds:
cmd = i.lower()
if cmd == 'solve-method':
if optcmds[i].lower() in list_of_solvers:
solver_cmd = optcmds[i].lower()
elif cmd == 'data-file':
problem_file_directory = os.path.dirname(os.path.realpath(__file__)) + '/Problem Files/'
### If the file is in the problem files folder then the filename is the sub-directory plus the filename
if os.path.isfile(os.path.join(problem_file_directory, optcmds[i])):
import_file_name = os.path.join(problem_file_directory,optcmds[i])
### Otherwise assume the file is in the current directory
else:
import_file_name = optcmds[i]
### file name for output directory
file_name = optcmds[i]
elif cmd == 'solver':
solver = optcmds[i].lower()
elif cmd == 'mipgap':
try:
mipgap = float(optcmds[i])
except:
raise Exception("MIP Gap must be a number")
elif cmd == 'none':
for ntms in optcmds[cmd]:
if ntms.lower() in solver_options:
_opts.append(ntms)
else:
raise Exception("Option " + str(i) + " Not Supported")
exit()
### Problem Data
try:
file_reader = ['import', import_file_name]
except:
raise Exception("Must Specify File")
### Set defaults
try:
solver
except:
solver = 'cplex'
try:
mipgap
except:
mipgap = .001
### Set output directory
current_directory = os.path.dirname(os.path.realpath(__file__))
current_date = time.strftime('%m_%d_%Y', time.gmtime())
if len(_opts) == 0:
output_directory = current_directory + '/Solutions/' + str(file_name) + '/' + str(solver_cmd) + '_' + current_date + '/'
else:
output_directory = current_directory + '/Solutions/' + str(file_name) + '/' + str(solver_cmd) + '_' + current_date + '_'+ str(_opts) + '/'
### Import Data
print("Importing data from " + str(file_name))
model_data = import_data_class.Data_Collection(file_reader)
### Import Solver method based on solver type and generate solve class
if solver_cmd == 'kda':
import Core.Solvers.KDA.KDA_Solution_Class as Solve
### Check options for KDA
kda_mipgap = .001
Solution = Solve.KDA(model_data, solver, kda_mipgap, output_directory,_opts)
### Import Solution Evaluation
import Core.Solvers.KDA.Evaluate_KDA_PRDP as Evaluate_KDA
### Calculate Equivalent ENPV
results = Solution.output['results']
sp_realizations = Solution.output['sub_problem_realizations']
Evaluated_Solution = Evaluate_KDA.KDA_PRDP_results(model_data,results,sp_realizations, output_directory)
### Write results to Consolidated file
import Core.output_write as output_write
output_write._write(Evaluated_Solution,Solution,output_directory,solver_cmd, file_name)
### If option exists, solve using initializations
if 'deterministic_initializations' in _opts:
### Send Results to Deterministic model
from Core.Solvers.KDA.KDA_Deterministic_Initializations import det_inits
print("Starting Deterministic Solve")
model_type = model_data._data['model_type'][None][0]
warmstart_output_directory = output_directory + '/Warmstart/'
det_inits(Solution, model_type, model_data, warmstart_output_directory, mipgap)
elif solver_cmd == 'mtssp':
import Core.Solvers.MTSSP.MTSSP_Solution_Class as Solve
### Check MTSSP Options
Solution = Solve.MTSSP(model_data,solver,mipgap,output_directory, _opts)
### Import Solution Evaluation
import Core.Evaluate_MTSSP as Evaluate_MTSSP
elif solver_cmd == 'mssp':
### Check for MSSP options
if len(_opts) == 0:
from Core.Solvers.MSSP.Deterministic_Solver import Deterministic_PRDP_Solve as Solve
Solution = Solve(mipgap,model_data, output_directory)
elif solver_cmd == 'saa' :
import Core.Solvers.SAA.SAA_Solution_Class as Solve
Solution = SAA(model_data)
elif solver_cmd == 'eoss' :
import Core.Solvers.EOSS.EOSS_Solution_Class as Solve
### Solve EOSS problem
Solution = Solve.EOSS(model_data)
return
