On the multi-product packing-delivery problem with a fixed route

We study a variation of the packing-delivery problem, which involves K distinct products, a single capacitated truck, and a fixed sequence of customer locations to be visited. We show that the non-split version of the problem is strongly polynomial solvable. For the split delivery case, we analyze the conditions of optimality and propose a local optimization algorithm with a computational complexity of O(n3 log(n)). Based on 14,000 randomly generated test problems, a consistent performance improvement achieved by the proposed local optimization algorithm over that by the optimal non-split packing-delivery schedules is observed.