Sensitivity analysis of dynamic lot-sizing heuristics

The deciding factors in making lot-sizing decisions are demand pattern, setup cost and holding cost. In reality, setup cost and holding cost are never really known. It is more reasonable to assume that these costs are known only within some given ranges. Therefore, it is important to understand the effects that parameter uncertainty may have on the inventory replenishment plans. This paper presents the sensitivity analysis of six well-known lot-sizing heuristics. Computational results show that the economic order quantity policy is the most sensitive method to errors in parameter estimation while the Silver-Meal heuristic is the most insensitive one. It is also concluded that the optimal solutions of dynamic lot-sizing problems are robust.