Extensions of simple eyeballing dynamic lot sizing heuristics

Abstract

A simple class of eyeballing heuristic algorithms for discrete dynamic lot size problems is presented with rolling horizon in which the forecast window is updated after each order cycle. These eye-balling heuristics are derived from the most direct and practical questions on inventory costs, and only require simple eyeballing comparisons of pairs of numbers representing the demands and some critical cut-off values. For more flexibility in coping with uncertainty, each critical cut-off value is estimated by an interval of lower and upper cut-off values. Simulation experiments are conducted and the results show that despite their extreme simplicity, these extensions of eyeballing heuristics have satisfactory performances compared with other heuristics. Furthermore, this simple-minded approach can be extended to the case of continuous time demands where heuristic solutions can be obtained by using simple graphical approximation. © 1998 Elsevier Science Ltd. All rights reserved.