An innovative approach for job pre-allocation to parallel unrelated machines in the case of a batch sequence dependent manufacturing environment

Abstract

The problem of allocating jobs to a set of parallel unrelated machines in a make to stock manufacturing system is studied. The items are subdivided into families of similar products. Sequence-dependent setups arise when products belonging both to the same family and to different families are sequenced. Restrictions on the number of available setups should be considered. The availability of planning batch production exists. Nevertheless, batch size is not known a priori. Hence, a solving approach considering both a pre-assignment procedure and a scheduling algorithm is proposed. Specifically, the focus of the article is on the pre-assignment methodology: a pre-assignment model (solved by a commercial solver) and two heuristics are presented and compared, in order to minimise the average idle residual capacity during the planning horizon, while considering pejorative factors related with the split volumes of the same product on different machines, unsatisfied demand along with demand produced in advance in each time period. The application to a case study is finally described in order to assess the performance of the proposed approach.

Keywords:

• scheduling
• parallel unrelated machines
• setup sequence dependent
• batch sizing