On heuristics and their performance evaluation for dynamic lot sizing (Q1117822)

The worst case performance ratios (WCPRs) of several heuristics that may be used as surrogates for the optimal dynamic lot sizing model of Wagner and Whitin were analyzed by several researchers. We propose one equally reasonable heuristic and present a simpler proof of the WCPR of the least unit cost heuristic. Since the WCPR is not only overly pessimistic but may also be meaningless, we propose in its place the concept of anticipated performance ratio that is based on the expected characteristics of the parameters of the problem over the planning horizon, and illustrate its application in the case of bounded demand.