Hi Suppose there is a 3PL which is in charge of delivery of products to many customers (more than 400 customers) every time with some specific constraints which completely differentiates it from conventional VRP problems. Exact solution approaches such as B&B, B&P, and B&C are not capable of solving this problem because of high complexity although they can handle all constraints. On the other hand, meta-heuristic algorithms seem to be appropriate. However, it might result in high computational time which is not desirable in real-world problems and even infeasibility because of a vast number of constraints. I would be thankful to know how this decision should be optimized.
asked
Amin-Sh |

Are you saying that exact methods (B&B etc.) cannot find "good" feasible solutions in acceptable time, or just that they cannot reach proven optimality? Since you are willing to entertain heuristics (which virtually never come with guarantee of optimality), a fair comparison would be to what an exact method can do if it focuses on objective value rather than bound.

Dear Rubin. Suppose you need to find a relatively good solution within 10 minutes for a VRP with 200 nodes or even more. Is there any exact algorithm that can find a good solution in a short time? In general, is there any approach to do this, no mather it is exact, heuristic, meta-heuristic?

Sorry, that question is too vague/general to give a meaningful answer. It depends on the specifics of the model, the capabilities of the computer, the software being used and the definition of "relatively good" ... and also depends on chance. It's certainly possible the answer is yes, and likely not guaranteed the answer is yes. All I can suggest is trying different things until one works.