CPLEX ignores a Initial feasible solution

Specific Domains Optimization (Mathematical)
1

Dear all,
I am solving a MILP with CPLEX 22.1. I have multiple objectives, and I minimize z1 first. I save the binary variables (called y[j,ℓ,t,τ]). Let Y0 the matriz with all y’s produced in the first stage.
I the second stage, I minimize z2, forcing z1 <= z1* + 0.1, by inserting Y0 as incumbent (only the binaries), to reduce the effort to solve the second problem. But, CPLEX ignores my solution (and it is feasible!).
When I fix these values at start value, I have a feasible problem.
I have used the following code to do this:

 for (j,ℓ,t,τ) in eachindex(y)
        set_start_value(y[j,ℓ,t,τ],Int(round(Y[j,ℓ,t,τ])))
 end

Anyone can help me please?

2

Hi @angeloaliano1,

Do you have the CPLEX log? How do you know that CPLEX ignored your solution?

Are there other discrete variables other than those binaries?

Could you try setting a primal start for all primal variables?