A MILP model for simultaneous supply chain and facility design considering production scheduling

FUMERO, YANINA; MONTAGNA, JORGE MARCELO; CORSANO, GABRIELA

Río de Janeiro

Conferencia; 3rd International Conference on Engineering Optimization – EngOpt 2012.; 2012

Graduate School in Engineering, Federal University of Rio de Janeiro

Several authors have addressed to the integration of supply chain (SC) decisions as an important and still open issue. The importance of interactions among decision levels is that significant benefits can be obtained by approaching the network as a whole and considering its various components simultaneously. Despite the growing body of literature about SC and related decisions (design, planning, scheduling, etc.), to the best of our knowledge, an integrated SC and plant design considering production scheduling remains to be addressed. In this work, a new formulation for the simultaneous design of SC and involved plants is presented, where the scheduling is specially considered through the use of mixed product campaigns (MPC). Besides providing a more steady supply of products from the commercial point of view, the sequencing of batches of different products in a MPC can reduce idle times and improve the utilization of the equipment. This approach is especially valid for stable contexts. In particular, this formulation is focused on the case of multiproduct batch plants. The SC considered in this work comprises three echelons: suppliers, multiproduct batch plants, and customers. Decisions regarding SC network such as nodes selection, supplier selection, material flows among nodes, and products distribution are jointly considered with multiproduct plant design (out of phase unit duplication and unit sizes), mixed product campaigns composition (number of batches of each product), and their sequencing in order to meet a specified economic criterion fulfilling demand requirements. With the aim of avoiding non linear formulations, some model assumptions are posed: discrete sizes for process units, maximum number of product batches in the campaign, and a set of discrete values for the number of campaign repetitions for each installed plant. The proposed model is applied to several examples where the different tradeoffs between plants design and scheduling and network design are assessed.