Optimisation approaches for supply chain planning and scheduling under demand uncertainty

This work presents efficient MILP-based approaches for the planning and scheduling of multiproduct multistage continuous plants with sequence-dependent changeovers in a supply chain network under demand uncertainty and price elasticity of demand. This problem considers multiproduct plants, where several products must be produced and delivered to supply the distribution centres (DCs), while DCs are in charge of storing and delivering these products to the final markets to be sold. A hybrid discrete/continuous model is proposed for this problem by using the ideas of the Travelling Salesman Problem (TSP) and global precedence representation. In order to deal with the uncertainty, we proposed a Hierarchical Model Predictive Control (HMPC) approach for this particular problem. Despite of its efficiency, the final solution reported still could be far from the global optimum. Due to this, Local Search (LS) algorithms are developed to improve the solution of HMPC by rescheduling successive products in the current schedule. The effectiveness of the proposed solution techniques is demonstrated by solving a large-scale instance and comparing the solution with the original MPC and a classic Cutting Plane approach adapted for this work.