This paper describes two new meta-heuristics which generate globally feasible, near optimal distribution system design. The first procedure gradually cools the system, but adapts a memory structure from the well-known tabu search. The second meta-heuristic allows for gradual cooling and sudden reheating of the system to avoid locally optimal solutions. Our results suggest that integrating simulated annealing with tabu search is recommended for this supply chain problem.
展开▼