the result of applying yalmip to solve MIP problem

[Reley]

Hello,professor

  I am trying to use yalmip to call the cplex to solve my unit commitment problem. There are no errors in running the code, but the final result is as follows, and the objective function has no value(it appears Linear scalar (real, 1251 variables)).Can you tell me why? Thanks!

* Starting YALMIP integer branch & bound.

* Lower solver   : fmincon-standard

* Upper solver   : rounder

* Max iterations : 300

 

Warning : The continuous relaxation may be nonconvex. This means 

that the branching process is not guaranteed to find a

globally optimal solution, since the lower bound can be

invalid. Hence, do not trust the bound or the gap...

 Node       Upper       Gap(%)      Lower    Open

[Johan Löfberg]

1. Cplex is not invoked. You have not specified any solver (or specified but never used options structure), as this is the display from the internal solver bnb

2. The fact that bnb is selected means that cplex is not applicable, and bnb is telling you why (relaxed problem nonconvex)

3. Most likely you have performed the cardinal sin of multiplying binary and continuous variable to model on/off. That's precisely what you learn in the unit commitment example that you never ever do https://yalmip.github.io/example/unitcommitment/

4. bnb as crashed, otherwise you would have seen iteration displays (and there should be an error code in the output from optimizer). Turn on 'debug' to see why it crashes

[Reley]

1.I specified the cplex solver as you suggested,like this,ops=sdpsettings('solver','cplex','verbos',2);output = optimize(ST,obj,ops);

2.When running code,I get results as follows:

Warning: Solver not applicable (cplex)

3.When I type cplex on the matlab command line, I get the following results. Does this mean that  cplex installation was successful, but it can't solve my unit commitment problem?

[Johan Löfberg]

I already told you, (2) and (3).