Updating the number of variables and Constraints through For Loop
AnsweredGood afternoon!
I'm using Gurobi in Python and I was hoping to update the number of variables using a for loop. For example, this is how I initially defined yDHT:
yDHT = m.addVars(D, H, S, vtype=GRB.CONTINUOUS, name='yDHT')
where S is a list containing 3 scenarios
Now, I want to increase the number of scenarios to 10 by using a for loop. I also want to solve the optimization model each time a scenario is added. This way, I'll have the initial run of the model containing 3 scenarios + 7 more runs of the models with the updated list of scenarios. For example
Initial optimization run: [S1, S2, S8]
Run optimization with the first scenario added: [S1, S2, S8, S3], while solutions from the previous run being the warm-start
Run optimization with the second scenario added: [S1, S2, S8, S3, S5], while solutions from the previous run being the warm-start
Run optimization with the third scenario added: [S1, S2, S8, S3, S5, S7], while solutions from the previous run being the warm-start
This is how I have done so far:
for scenario_update in range(len(scenarios1.Index)):
entering_scenario = scenarios1.take([scenario_update])
scenarios = scenarios.append(entering_scenario)
scenario = scenarios.Index.tolist()
Which returns the right sequence of scenarios:
[S1, S2, S8] -> [S1, S2, S8, S3] -> [S1, S2, S8, S3, S5] -> [S1, S2, S8, S3, S5, S7]
But I wasn't sure how I should go about updating the constarints for each of the scenarios added, and updating the number of variables.
Should I do something like this?
#Updating the number of variables
YDHT.update({depot, entering_scenario, m.addVar(vtype = GRB.BINARY, name= 'yDHT')})
#Updating Constraints
for s in entering_scenario.Index.tolist():
for d in depot:
m.addConstr(gp.quicksum(yDBT[d, b, s] for b in DC) +
gp.quicksum(yDHT[d, h, s] + yDHV[d, h, s] + yDHU[d, h, s] for h in shelter)
<= PD[d] * xD[d], name='Constraint 4')
#Warm-starting
yDHT[depot[d], shelter[h], scenario[s]].start = yDHT[depot[d], shelter[h], scenario[s]].x
#Re-run the model
m.optimize(callback = data_cb)
Or is there a better approach?
Thank you so much. Any help will be truly appreciated.
Sincerely,
Min Lee
-
Official comment
This post is more than two years old. Some information may not be up to date. For current information, please check the Gurobi Documentation or Knowledge Base. If you need more help, please create a new post in the community forum, or try Gurobot, our chatbot interface offering instant, expert-level support. -
Hi Min Lee,
I think that it is easier to have a list or a dictionary holding multiple yDHT entries instead of directly modifying the tupledict returned by the addVars method.
You could have a dictionary similar to
S = {S1: "yHDT defined over scenario S1", S2: "yHDT defined over scenario S2", ...}Then, when adding constraints, you can run over all entries in this dictionary and construct the corresponding constraints.
Best regards,
Jaromił0
Post is closed for comments.
Comments
2 comments