Using metaheuristics for solving a production scheduling problem in a chemical firm. A case study

Abstract The problem is raised by a workshop management team in a chemical industry producing materials of different types. The workshop contains several equipments in parallel and in series which can be used for the production of each job. The equipments in series are connected by a piping system forming a network by which the material runs out. Some connections are shared by several pairs of equipments so that the use of such a connection for one pair of equipments makes this connection unavailable at the same time for other pairs. Several particular constraints must be satisfied and deadlines exist for some jobs. The objectives are to minimize the makespan and the tardiness. A method is proposed to heuristically obtain a satisfying production schedule. The procedure involves two stages. In a first step, given a fixed order of the jobs, a conventional heuristic is proposed to successively assign the jobs to the available equipments. In a second step, the simulated annealing or tabu metaheuristics are applied to optimize the order of the jobs. The method was implemented on a PC and applied to an instance of the problem. Some comparisons are made to analyse and compare the efficiency of the heuristics.