Lot streaming for customer order scheduling problem in job shop environments

Abstract

In this study, several jobs make up a given customer order, and each job consists of a batch size greater than one, and nothing is delivered to the customer until the order is complete. Until delivery can be made, the completed jobs are to be stored in the inventory. In an order-based environment, scheduling is usually referred to as a customer order scheduling (COS) problem. The lot streaming (LS) technique is to split a processing job into several sub-jobs such that successive operations of the same job can be overlapped. LS has been shown to be an effective technique for compressing the mean lead time in a job-based environment, but has not been studied in the order-based environment. Therefore, this paper extends, for the first time, the applicability of LS to the COS problem, in order to investigate whether the expected benefits of LS can be observed in an order-based environment. To solve this complex problem, a genetic algorithm (GA) is applied to determine the LS conditions. The experiments led to the conclusion that our proposed algorithm (LS-GA) is significantly superior over the other LS modes in terms of makespan, lateness and finished goods (FG) flow time.