Originally posted by: WangFISH
Hi
I get some problems in using the Cplex to solve my model. The Cplex is already running for 20 hours for this model and still cannot solve it .
I find the key of this probelem is one of my constraint in the model: " forall(a in A,b in B,c in C,t in T) Z[a][b][c][t]>=X[a][t]+K[a][t]*Y[a][b][c]-K[a][t];" where Z and X is decision variables, Y is binary variable and K is known as 'K=[[492,195],[449,208],[447,212],[444,198],[462,239]];'
I find that :
1. If I delete this constraint, the Cplex can solve the model in 10s.
2.If I change the last term of this constraint as :" forall(a in A,b in B,c in C,t in T) Z[a][b][c][t]>=X[a][t]+K[a][t]*Y[a][b][c]-1000;" /*or some value far more than max(K)=492) */ ,the Cplex can solve the model in a short time.
3.If I change the last term of this constraint as :" forall(a in A,b in B,c in C,t in T) Z[a][b][c][t]>=X[a][t]+K[a][t]*Y[a][b][c]-495;" /*or some value which is more than 492 and also close to 492 */ the Cplex will keep running this model for a long long time and cannot solve it.
4.If I change the last term of this constraint as :" forall(a in A,b in B,c in C,t in T) Z[a][b][c][t]>=X[a][t]+K[a][t]*Y[a][b][c]-400;" /*or some value between 369 to 491 */ the Cplex can solve the model but every decision variable is Null.
5.If I change the last term of this constraint as:" forall(a in A,b in B,c in C,t in T) Z[a][b][c][t]>=X[a][t]+K[a][t]*Y[a][b][c]-0;" /*or some value less than 368 */ It will show that "CP Optimizer solves problems with a search space of up to 2^1000" which I think is casued by a contradiction with other constraints.
Really hope you can help me.
Sincerely
#DecisionOptimization#OPLusingCPOptimizer