How to set the timeout for Pyomo solve() method ? More specifically, to tell pyomo, after x seconds, return the optimal solution currently found ?
Asked
Active
Viewed 9,019 times
2 Answers
12
So I was able to find the answer via pyomo documentation and I thought it would be helpful to share.
To set the timeout for Pyomo solve() method:
solver.solve(model, timelimit=5)
However this will throw pyutilib.common._exceptions.ApplicationError: "Solver (%s) did not exit normally" % self.name ) if the solver is not terminated. What I really want is to pass the timelimit option to my solver. In my case of cplex solver, the code will be like this:
solver = SolverFactory('cplex')
solver.options['timelimit'] = 5
results = solver.solve(model, tee=True)
More on pyomo and cplex docs.
zyzo
- 305
- 3
- 12
-
2I am doing the same thing (although using 'cbc' solver, not sure if that is relevant or not) but I don't think this answers the original question: "tell pyomo, after x seconds, return the optimal solution currently found." This just seems to throw the error mentioned. Should it return an actual "solution" even if it's not optimal and/or doesn't satisfy all the constraints? Is that possible? – seth127 Sep 24 '19 at 18:38
8
I had success with the following in Pyomo. The name of the time limit option is different for different solvers:
self.solver = pyomo.opt.SolverFactory(SOLVER_NAME)
if 'cplex' in SOLVER_NAME:
self.solver.options['timelimit'] = TIME_LIMIT
elif 'glpk' in SOLVER_NAME:
self.solver.options['tmlim'] = TIME_LIMIT
elif 'gurobi' in SOLVER_NAME:
self.solver.options['TimeLimit'] = TIME_LIMIT
elif 'xpress' in SOLVER_NAME:
self.solver.options['soltimelimit'] = TIME_LIMIT
# Use the below instead for XPRESS versions before 9.0
# self.solver.options['maxtime'] = TIME_LIMIT
Where TIME_LIMIT is an integer time limit in seconds.
