Multi-objective Evolutionary Scheduling of Distributed Supply Networks

This paper considers the problem of scheduling a distributed network of production centers supplying a quickly perishable good that has to be produced just-intime and delivered within customer-specified strict time windows. The problem includes several planning, scheduling and routing problem, each notoriously affected by nearly prohibitive combinatorial complexity. Ideal solutions should provide a good compromise between production costs (resource utilization, delivery costs) and tolerance to stochastic perturbations (transport delays). We propose a novel multiobjective meta-heuristic approach based on a hybrid genetic algorithm combined with constructive heuristics. The algorithm is designed to return a set of solutions with different cost and risk tradeoffs. The effectiveness of the approach is confirmed by a comparison with other recently proposed methods. Copyright 2005 IFAC