A scheduling algorithm for the reentrant shop: an application in semiconductor manufacture

Abstract

This paper addresses a heuristic algorithm that minimizes total weighted tardiness for reentrant flow shop problems with sequence-dependent set up time. The proposed algorithm consists of three phases: prior plan phase, initial solution generation phase and solution improvement phase. In the first phase, a reentrant order with cyclic process is divided into a number of jobs. An initial solution is obtained by a RATCS (revised apparent tardiness cost with setups) rule in the second phase, and the last phase employs a RHTS (rolling horizon tabu search) to improve the solution. The experimental results show that the proposed algorithms give very efficient schedules in terms of total weighted tardiness and computational effort.