def _test_root_infeasible_min(comm):
solver = Solver(comm=comm)
baseline1 = SolverResults()
baseline1.solution_status = "infeasible"
baseline1.termination_condition = "optimality"
baseline1.objective = inf
baseline1.bound = inf
baseline1.nodes = 1
baseline1.best_node = None
baseline1.wall_time = _ignore_value_
baseline2 = SolverResults()
baseline2.solution_status = "infeasible"
baseline2.termination_condition = "optimality"
baseline2.objective = inf
baseline2.bound = inf
baseline2.nodes = 0
baseline2.best_node = None
baseline2.wall_time = _ignore_value_
problem = root_infeasible_min.RootInfeasibleMin()
for queue_strategy in _queue_strategies:
if queue_strategy == "custom":
continue
results = _execute_single_test(problem,
baseline1,
solver=solver,
queue_strategy=queue_strategy)
queue = solver.save_dispatcher_queue()
_execute_single_test(problem,
baseline2,
solver=solver,
queue_strategy=queue_strategy,
initialize_queue=queue,
best_objective=results.objective,
best_node=results.best_node)
def _test_delayed_unbounded_min(comm):
solver = Solver(comm=comm)
baseline1 = SolverResults()
baseline1.solution_status = "unbounded"
baseline1.termination_condition = "optimality"
baseline1.objective = -inf
baseline1.bound = -inf
baseline1.nodes = _ignore_value_
baseline1.best_node = None
baseline1.wall_time = _ignore_value_
baseline2 = SolverResults()
baseline2.solution_status = "unbounded"
baseline2.termination_condition = "optimality"
baseline2.objective = -inf
baseline2.bound = -inf
baseline2.nodes = 0
baseline2.best_node = None
baseline2.wall_time = _ignore_value_
problem = delayed_unbounded_min.DelayedUnboundedMin()
for queue_strategy in _queue_strategies:
if queue_strategy == "custom":
continue
results = _execute_single_test(problem,
baseline1,
solver=solver,
queue_strategy=queue_strategy)
queue = solver.save_dispatcher_queue()
_execute_single_test(problem,
baseline2,
solver=solver,
queue_strategy=queue_strategy,
initialize_queue=queue,
best_objective=results.objective,
best_node=results.best_node)
# test the NestedSolver
results = _execute_single_test(NestedSolver(problem,
time_limit=None,
node_limit=None),
baseline1,
solver=solver)
queue = solver.save_dispatcher_queue()
_execute_single_test(NestedSolver(problem,
time_limit=None,
node_limit=None),
baseline2,
solver=solver,
initialize_queue=queue,
best_objective=results.objective,
best_node=results.best_node)
def _execute_single_test(problem, baseline, solver=None, comm=None, **kwds):
if solver is None:
solver = Solver(comm=comm)
else:
assert comm is None
if solver.comm is not None:
if solver.comm.rank == 0:
pass
elif solver.comm.rank == 1:
pass
elif solver.comm.rank == 3:
pass
orig = Node()
problem.save_state(orig)
results = solver.solve(problem, **kwds)
assert isinstance(results.solution_status, six.string_types)
assert isinstance(results.termination_condition, six.string_types)
current = Node()
problem.save_state(current)
assert current.state == orig.state
assert len(vars(results)) > 0
assert len(vars(results)) == len(vars(baseline))
for name in sorted(sorted(list(vars(results).keys()))):
if getattr(baseline, name) is _ignore_value_:
continue
if (name == 'nodes') and \
(solver.comm is not None) and \
(solver.comm.size > 2):
pass
else:
assert getattr(results, name) == getattr(baseline, name), \
("value for '"+str(name)+"' ("+
str(getattr(results, name))+") does "
"not match baseline ("+
str(getattr(baseline, name))+")")
if solver.is_dispatcher:
q = solver.save_dispatcher_queue()
assert len(q.nodes) == solver._disp.queue.size()
assert q.sense == solver._disp.converger.sense
assert q.worst_terminal_bound == solver._disp.worst_terminal_bound
assert q.bound() == results.bound
return results
def _execute_tests(comm, problem, baseline, **kwds):
assert 'log_interval_second' not in kwds
assert 'log' not in kwds
_execute_single_test(problem, baseline, comm=comm, **kwds)
kwds['log_interval_seconds'] = 0.0
_execute_single_test(problem, baseline, comm=comm, **kwds)
kwds['log_interval_seconds'] = 100.0
_execute_single_test(problem, baseline, comm=comm, **kwds)
kwds['log_interval_seconds'] = 0.0
kwds['log'] = None
_execute_single_test(problem, baseline, comm=comm, **kwds)
kwds['log_interval_seconds'] = 100.0
_execute_single_test(problem, baseline, comm=comm, **kwds)
solver = Solver(comm=comm)
_execute_single_test(problem, baseline, solver=solver, **kwds)
_execute_single_test(problem, baseline, solver=solver, **kwds)
kwds['log'] = get_simple_logger(level="WARNING")
_execute_single_test(problem, baseline, solver=solver, **kwds)
def _test_infeasible_max(comm):
solver = None
if comm is not None:
solver = Solver(comm=comm)
baseline = SolverResults()
baseline.solution_status = "infeasible"
baseline.termination_condition = "optimality"
baseline.objective = -inf
baseline.bound = -inf
baseline.nodes = 255
baseline.best_node = None
baseline.wall_time = _ignore_value_
problem = infeasible_max.InfeasibleMax()
if comm is None:
_execute_tests(comm, problem, baseline)
else:
_execute_single_test(problem, baseline, solver=solver)
baseline = SolverResults()
baseline.solution_status = "unknown"
baseline.termination_condition = "queue_empty"
baseline.objective = -inf
baseline.bound = -16
baseline.nodes = 31
baseline.best_node = None
baseline.wall_time = _ignore_value_
problem = infeasible_max.InfeasibleMax(branching_abstol=0.1)
if comm is None:
_execute_tests(comm, problem, baseline)
else:
_execute_single_test(problem, baseline, solver=solver)
baseline = SolverResults()
baseline.solution_status = "unknown"
baseline.termination_condition = "objective_limit"
baseline.objective = -inf
baseline.bound = -16
baseline.nodes = 31
baseline.best_node = None
baseline.wall_time = _ignore_value_
problem = infeasible_max.InfeasibleMax(branching_abstol=0.1)
if comm is None:
_execute_tests(comm, problem, baseline, bound_stop=-15)
else:
_execute_single_test(problem, baseline, solver=solver, bound_stop=-15)
baseline = SolverResults()
baseline.solution_status = "feasible"
baseline.termination_condition = "objective_limit"
baseline.objective = -20
baseline.bound = -16
baseline.absolute_gap = 4
baseline.relative_gap = 0.2
baseline.nodes = 31
baseline.best_node = _ignore_value_
baseline.wall_time = _ignore_value_
problem = infeasible_max.InfeasibleMax(branching_abstol=0.1,
fixed_objective=-20)
if comm is None:
_execute_tests(comm, problem, baseline, bound_stop=-15)
else:
_execute_single_test(problem, baseline, solver=solver, bound_stop=-15)
baseline = SolverResults()
baseline.solution_status = "unknown"
baseline.termination_condition = "node_limit"
baseline.objective = -inf
baseline.bound = -16
baseline.nodes = 31
baseline.best_node = None
baseline.wall_time = _ignore_value_
problem = infeasible_max.InfeasibleMax()
if comm is None:
_execute_tests(comm, problem, baseline, node_limit=31)
elif comm.size <= 2:
# skip for larger comm sizes
# as the node_limit can lead to
# a number of different outcomes
_execute_single_test(problem, baseline, solver=solver, node_limit=31)
baseline = SolverResults()
baseline.solution_status = "feasible"
baseline.termination_condition = "node_limit"
baseline.objective = -17
baseline.bound = -16
baseline.absolute_gap = 1.0
baseline.relative_gap = 1.0 / 17
baseline.nodes = 31
baseline.best_node = _ignore_value_
baseline.wall_time = _ignore_value_
problem = infeasible_max.InfeasibleMax(fixed_objective=-17)
if comm is None:
_execute_tests(comm, problem, baseline, node_limit=31)
elif comm.size <= 2:
# skip for larger comm sizes
# as the node_limit can lead to
# a number of different outcomes
_execute_single_test(problem, baseline, solver=solver, node_limit=31)
baseline = SolverResults()
baseline.solution_status = "unknown"
baseline.termination_condition = "time_limit"
baseline.objective = -inf
baseline.bound = inf
baseline.nodes = 0
baseline.best_node = None
baseline.wall_time = _ignore_value_
problem = infeasible_max.InfeasibleMax()
if comm is None:
_execute_tests(comm, problem, baseline, time_limit=0)
else:
_execute_single_test(problem, baseline, solver=solver, time_limit=0)
if solver is None:
solver = Solver(comm=comm)
baseline = SolverResults()
baseline.solution_status = "infeasible"
baseline.termination_condition = "optimality"
baseline.objective = -inf
baseline.bound = -inf
baseline.nodes = 255
baseline.best_node = None
baseline.wall_time = _ignore_value_
problem = infeasible_max.InfeasibleMax()
for queue_strategy in _queue_strategies:
if queue_strategy == "custom":
continue
_execute_single_test(problem,
baseline,
solver=solver,
queue_strategy=queue_strategy)
baseline = SolverResults()
baseline.solution_status = "invalid"
baseline.termination_condition = "optimality"
baseline.objective = -100000000
baseline.bound = -inf
baseline.nodes = _ignore_value_
baseline.best_node = None
baseline.wall_time = _ignore_value_
problem = infeasible_max.InfeasibleMax()
for queue_strategy in _queue_strategies:
if queue_strategy == "custom":
continue
_execute_single_test(problem,
baseline,
solver=solver,
queue_strategy=queue_strategy,
best_objective=-100000000)
baseline1 = SolverResults()
baseline1.solution_status = "infeasible"
baseline1.termination_condition = "optimality"
baseline1.objective = -inf
baseline1.bound = -inf
baseline1.nodes = _ignore_value_
baseline1.best_node = None
baseline1.wall_time = _ignore_value_
baseline2 = SolverResults()
baseline2.solution_status = "infeasible"
baseline2.termination_condition = "optimality"
baseline2.objective = -inf
baseline2.bound = -inf
baseline2.nodes = _ignore_value_
baseline2.best_node = None
baseline2.wall_time = _ignore_value_
problem = infeasible_max.InfeasibleMax()
for queue_strategy in _queue_strategies:
if queue_strategy == "custom":
continue
results = _execute_single_test(problem,
baseline1,
solver=solver,
queue_strategy=queue_strategy)
queue = solver.save_dispatcher_queue()
_execute_single_test(problem,
baseline2,
solver=solver,
queue_strategy=queue_strategy,
initialize_queue=queue,
best_objective=results.objective,
best_node=results.best_node)
def _test_zero_objective_min(comm):
solver = Solver(comm=comm)
baseline = SolverResults()
baseline.solution_status = "optimal"
baseline.termination_condition = "optimality"
baseline.objective = 0.0
baseline.bound = -0.0078125
baseline.absolute_gap = 0.0078125
baseline.relative_gap = 0.0078125
baseline.nodes = 255
baseline.best_node = _ignore_value_
baseline.wall_time = _ignore_value_
problem = zero_objective_min.ZeroObjectiveMin()
_execute_single_test(problem,
baseline,
solver=solver,
relative_gap=0.01,
absolute_gap=0.01)
_execute_single_test(problem,
baseline,
solver=solver,
relative_gap=0.0,
absolute_gap=0.01)
_execute_single_test(problem,
baseline,
solver=solver,
relative_gap=0.01,
absolute_gap=0.0)
baseline = SolverResults()
baseline.solution_status = "optimal"
baseline.termination_condition = "optimality"
baseline.objective = 0.0
baseline.bound = -0.0009765625
baseline.absolute_gap = 0.0009765625
baseline.relative_gap = 0.0009765625
baseline.nodes = 2047
baseline.best_node = _ignore_value_
baseline.wall_time = _ignore_value_
problem = zero_objective_min.ZeroObjectiveMin()
_execute_single_test(problem,
baseline,
solver=solver,
relative_gap=0.001,
absolute_gap=0.001)
_execute_single_test(problem,
baseline,
solver=solver,
relative_gap=0.0,
absolute_gap=0.001)
_execute_single_test(problem,
baseline,
solver=solver,
relative_gap=0.001,
absolute_gap=0.0)
baseline = SolverResults()
baseline.solution_status = "feasible"
baseline.termination_condition = "queue_empty"
baseline.objective = 0.0
baseline.bound = -0.0009765625
baseline.absolute_gap = 0.0009765625
baseline.relative_gap = 0.0009765625
baseline.nodes = 2047
baseline.best_node = _ignore_value_
baseline.wall_time = _ignore_value_
problem = zero_objective_min.ZeroObjectiveMin()
_execute_single_test(problem,
baseline,
solver=solver,
relative_gap=0.0001,
absolute_gap=0.0001)
_execute_single_test(problem,
baseline,
solver=solver,
relative_gap=0.0,
absolute_gap=0.0001)
_execute_single_test(problem,
baseline,
solver=solver,
relative_gap=0.0001,
absolute_gap=0.0)
baseline = SolverResults()
baseline.solution_status = "optimal"
baseline.termination_condition = "optimality"
baseline.objective = 0.0
baseline.bound = -0.0078125
baseline.absolute_gap = 0.0078125
baseline.relative_gap = 0.0078125
baseline.nodes = None
baseline.best_node = _ignore_value_
baseline.wall_time = _ignore_value_
problem = zero_objective_min.ZeroObjectiveMin()
for queue_strategy in _queue_strategies:
if queue_strategy == "custom":
continue
elif queue_strategy == "random":
baseline.nodes = _ignore_value_
else:
baseline.nodes = 255
_execute_single_test(problem,
baseline,
solver=solver,
relative_gap=0.01,
absolute_gap=0.01,
queue_strategy=queue_strategy)