A mixed-integer optimization model for determining plant location in an integrated petrochemical industry

Abstract Potential investment in the Brazilian petrochemical industry over the next few years, due to a recently approved tax incentive program, raises the problem of industry planning. This work presents a multiperiod, mixed-integer optimization model for plant location in the Brazilian petrochemical industry. The objective function to be minimized is the total cost of the industry. The set of decision variables includes the location of new process units and the process capacities over 20 years. The model considers 136 products, 112 processes, and 24 locations. The solution indicates that some process units, such as polyether polyol and polyurethane units, are installed in locations with low investment and operational costs to meet market demand in other regions. Conversely, capacity expansions at polyethylene terephthalate and polyvinyl chloride plants occur in locations with considerable demand. A sensitivity analysis shows that the former strategy may be adopted when transportation costs are low relative to other costs. Additionally, processes for producing dimethyl terephthalate, epoxy resin, Nylon 6, polyacrylate, polyisobutylenes, polyether polyol, polyurethane, and acrylic fibers are consistently selected for investment, despite large variations in transportation costs. These results may support government and companies’ policies on the long-term planning of the Brazilian petrochemical industry.