A hybrid genetic algorithm for a multilevel assembly replenishment planning problem with stochastic lead times. (November 2020)

• Record Type: Journal Article
• Authors: Ben-Ammar, Oussama; Castagliola, Philippe; Dolgui, Alexandre; Hnaien, Faicel
• Is Part Of: Computers & industrial engineering. Volume 149(2020)
• Publication Date: 2020-11
• Subjects: Assembly systems -- Assemble-to-order -- Replenishment planning -- Stochastic lead-times -- Genetic algorithms -- Discrete optimization; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2020.106794
• Languages: English
• ISSNs: 0360-8352

Abstract

Abstract: This paper on replenishment planning for multi-level assembly systems with several components at each level deals with the problem of calculating planned lead-times when the real lead-times for all components are assumed to be stochastic. This problem is already treated in the literature by using a recursive procedure and a Branch and Bound algorithm. Here, in order to decrease the computation time, a novel generalized probabilistic model based on an iterative approach is developed. The proposed model calculates the expected total cost, which is composed of the inventory holding cost for components and the backlogging and inventory holding costs for the finished product. An iterative approach and a hybrid genetic algorithm are introduced to determine the planned order release dates for components at the last level of the bill of materials that minimizes the expected total cost. Experimental results show that the proposed optimization algorithm efficiently finds good-quality approximate solutions regardless of the type of assembly system, the number of components at the last level and the variability of the finished product-related costs. Highlights: Studies multi-level assembly systems with a fixed finished product demand and stochastic component lead times. Develops a generalization of the models introduced in several studies. Proposes a new efficient mathematical model. Develops an efficient algorithm to resolve large problems.