def run(self):
"""
This method execute the optimization on the knapsack problem.
@ In, None
@ Out, None
"""
outputDict = {}
if self.uncertainties is None:
inputData = self.generateModelInputData()
# specifying the path to a solver
# with SolverFactory(self.solver, executable=self.executable) as opt:
with SolverFactory(self.solver) as opt:
opt.options.update(self.sopts) # add solver options
model = self.createInstance(inputData)
results = opt.solve(model, load_solutions=False, tee=self.tee, **{'use_signal_handling':False})
if results.solver.termination_condition != TerminationCondition.optimal:
raise RuntimeError("Solver did not report optimality:\n%s" %(results.solver))
model.solutions.load_from(results)
outputDict.update(self.printSolution(model))
self.output.update(outputDict)
# TODO: Add collect output and return a dictionary for raven to retrieve information
else:
tree_model = self.pysp_scenario_tree_model_callback()
if self.stochSolver == 'ef':
# fsfct the callback function for pysp_instance_creation_callback
# tree_model generated by pysp_scenario_tree_model_callback
stsolver = rapper.StochSolver("", fsfct=self.pysp_instance_creation_callback, tree_model=tree_model)
ef_sol = stsolver.solve_ef(self.solver, sopts=self.sopts, tee=self.tee)
if ef_sol.solver.termination_condition != TerminationCondition.optimal:
raise RuntimeError("Solver did not report optimality:\n%s" %(ef_sol.solver))
# TODO: Add collect output and return a dictionary for raven to retrieve information
self.printScenarioSolution(stsolver)
outputDict.update(self.getScenarioSolution(stsolver.scenario_tree))
self.output.update(outputDict)
elif self.stochSolver == 'ph':
# fsfct the callback function for pysp_instance_creation_callback
# tree_model generated by pysp_scenario_tree_model_callback
stsolver = rapper.StochSolver("", fsfct=self.pysp_instance_creation_callback, tree_model=tree_model, phopts=self.phopts)
ph_sol = stsolver.solve_ph(subsolver=self.solver, default_rho=self.phRho, phopts=self.phopts, sopts=self.sopts, tee=self.tee)
# TODO: Add collect output and return a dictionary for raven to retrieve information
# first retrieve the xhat solution from solver, then print the solution
obj, xhat = rapper.xhat_from_ph(ph_sol)
for nodeName, varName, varValue in rapper.xhat_walker(xhat):
print (nodeName, varName, varValue)
if ph_sol.solver.termination_condition != TerminationCondition.optimal:
raise RuntimeError("Solver did not report optimality:\n%s" %(ph_sol.solver))
self.printScenarioSolution(stsolver)
outputDict.update(self.getScenarioSolution(stsolver.scenario_tree))
self.output.update(outputDict)
return outputDict
def test_ph_solve(self):
""" use ph"""
phopts = {'--max-iterations': '2'}
stsolver = rapper.StochSolver("ReferenceModel.py",
tree_model = self.farmer_concrete_tree,
fsfct = "pysp_instance_creation_callback",
phopts = phopts)
ph = stsolver.solve_ph(subsolver = solvername, default_rho = 1,
phopts=phopts)
obj = stsolver.root_E_obj() # E[xbar]
obj, xhat = rapper.xhat_from_ph(ph)
for nodename, varname, varvalue in rapper.xhat_walker(xhat):
pass
assert(nodename == 'RootNode')
def test_ph_solve(self):
""" use ph; assumes concrete two-stage json passes"""
phopts = {'--max-iterations': '2'}
stsolver = rapper.StochSolver("ReferenceModel.py",
tree_model=self.farmer_concrete_tree,
phopts=phopts)
ph = stsolver.solve_ph(subsolver=solvername,
default_rho=1,
phopts=phopts)
obj = stsolver.root_E_obj() # E[xbar]
obj, xhat = rapper.xhat_from_ph(ph)
for nodename, varname, varvalue in rapper.xhat_walker(xhat):
pass
assert (nodename == 'RootNode')