Assuming that your problem is actually feasible, try the following.
- Make up a feasible solution (a set of routes that satisfy all constraints, including time windows). It does not matter how poor the objective value is, just that it is feasible.
- In your model, fix all the variables at the corresponding values (by setting lower bound = upper bound = value).
- Try to solve your modified model. Presumably CPLEX will declare it infeasible.
- As CPLEX for an irreducible infeasible subset (IIS). If your candidate solution really is feasible, at least one of the constraints in the IIS (other than the bounds you set) must contain an error.
------------------------------
Paul Rubin
Professor Emeritus
Michigan State University
------------------------------