Modeling and optimization for supply-chain design and operation in process industry

Abstract

Design and operation of supply chains in process industry involve the diversified products to meet the various demands of global customers as well as the layers of facilities located at different sites. An integrated planning model is proposed to support the strategic decisions of the production and distribution in the refineries at different locations. The network planning of individual refineries are incorporated into their production planning of multiple fuel products. In addition, the integrated model is further developed for the collaborative planning of multi-site refineries. The simulation results based on the model indicate that the distribution costs can be reduced by relocating distribution centers as well as by reconfiguring their linkages to various markets. Moreover, the optimization results of the integrated network of multi-site refineries reveal the expected advantages of the individual refineries from the collaboration or the merger. Modeling and optimization for supply-chain design and operation tend to be profoundly convoluted because of its combinatorial complexities. Thus, excessively long time and large efforts are often required for formulation and computation. A graph-theoretic approach using Process-graphs (P-graphs) is adapted for the design of supply-chain so as to overcome the difficulties with mixed integer programming or other NP-hard methods. Every functional unit which contributes to the change of the objective function is interpreted as an operating unit in process networks. The proposed method enables the detailed description of networks and provides the near-optimal solutions as well as the optimal ones, thereby allowing more involvement of decision makers in the optimization process. The usefulness of the proposed method is demonstrated with both a hypothetical and a realistic example of process industry.