1. ¹City University of Hong Kong, Hong Kong
2. ²The Hong Kong Polytechnic University, Kowloon, Hong Kong

Correspondence: Z Xu, Department of Logistics, Faculty of Business, The Hong Kong Polytechnic University, Kowloon, Hong Kong. E-mail: lgtzx@polyu.edu.hk

Received August 2006; Accepted August 2007; Published online 16 January 2008.

Abstract

The resequencing and feature assignment problem (RFAP) appears among jobs in the assembly line, especially in the automotive industry. Each job in the assembly line must be assigned a feature from its feasible feature set. However, a changeover cost is incurred between two consecutive jobs with different features. To minimize the total changeover cost, the job sequence needs to be rearranged, but the rearrangement is restricted to the number of offline buffers. The RFAP turns out to be -hard in the strong sense. Based on a beam search heuristic to generate upper bounds of optimum solutions, we have proposed an iterative search scheme which can achieve optimum solutions in a reasonably short time, for cases sized as large as that in reality. Extensive experiments have shown very favourable results for our methods in terms of both the solution quality and the time efficiency.