A new GRASP and path relinking for single machine scheduling with sequence dependent setups

Setup (changeover) time is often needed by the processes with operations of cleaning or changing fixtures. Its length strongly depends on the immediately succeeding operation sometimes. The scheduling problems with such setup times are called to have sequence dependent setups. This paper focuses on a single machine scheduling problem with sequence dependent setups to minimize total weighted tardiness. A greedy randomized adaptive search procedure (GRASP) combined with path relinking (PR) is proposed for the NP-hard problem. In the algorithm, our main contributions are a new iterated filter-and-fan (IFF) algorithm in the improvement phase of GRASP and a block insertion based path construction method in PR. We compare our algorithm with discrete differential evolution (DDE), discrete particle swarm optimization (DPSO) and single IFF algorithm on a set of test instances from literature. The results show that the algorithm performs very competitively.