A reactive GRASP and path relinking for a combined production-distribution problem
Computers and Operations Research
Hi-index | 0.00 |
We investigate multiproduct, multilocation production/distribution networks with deterministic, stationary demands. Most research on such systems assumes zero leadtimes. Methods using fixed reorder intervals that are power-of-two multiples of a base planning period have proved to be very successful for such systems. The same methods apply to networks with positive leadtimes, provided the leadtimes arebalanced in a specific sense. We explore networks with general,unbalanced leadtimes. A key result is a lower bound on the cost of any feasible policy. Other results concern policy heuristics and their performance. For a large class of networks, we construct a policy whose cost is within 45% of the lower bound. For general networks, the performance guarantee is 1:02v(1+?), where ? is a number dependent on the network topology only. In general, the best performance bound is obtained by systematically reducing the order intervals derived from the corresponding zero-leadtime system.