Date of Publication :31st May 2024

Abstract:In supply chain management, on time delivery is a very important issue for customers. The customer order includes more than one item at all times. How to finish the order on time, so that all items in the same order will be ready for delivery, is an important task. Therefore, this paper considers the dynamic demand lot-sizing problem (DLSP) and the customer-ordering problem (COP) together, namely DLSCOP. DLSP focuses on the deterministic time-varying batch ordering lot-sizing problem with backorders. The COP consists of a set of items that must be shipped as one batch at the same time. This work develops a Linear Programming model for describing the DLSCOP, and applies genetic algorithms (GA) to the problem. This study also modifies two popular algorithms for benchmarking, Silver-Meal (SM) algorithms, and Wagner-Whitin (WW) algorithms, and develops two heuristics, MSM and MWW for solving these kinds of problem. The simulation test considers 128 scenarios and 100 repetitions. In the statistical analysis, the GA performance is better than MSM and MWW. The decision based on GA saves more than 10-50% cost, especially in those scenarios with long term, multiple items, and high expense rate (ordering cost and holding cost).

Reference :