Minimising earliness and tardiness on parallel machines with sequence-dependent setups

In this paper, we present heuristics for minimising total earliness and tardiness on a group of identical, parallel machines subject to sequence-dependent setups. Each job has a potentially unique due date, processing time and/or ready time. Our approach generates an initial feasible schedule using new algorithms and a dispatching rule, and then employs a local search procedure to improve this initial schedule when possible. Experimental results demonstrate the viability of our approach over a wide range of problem instances to produce near-optimal solutions in practically acceptable computation times.