Metaheuristic approaches to the pooling problem

The pooling problem, which has several application areas in chemical industry, is an extension of the blending problem and aims to find the optimal composition of materials in a two-stage network while obeying quality limitations for the end products. The pooling problem has a bilinear structure and it is NP-hard. The exact methods to solve the pooling problem are inefficient for large instances and a few heuristic methods exist. In this thesis, our aim is to propose two metaheuristic methods that are based on particle swarm optimization (PSO) and simulated annealing (SA). Both of the proposed approaches take advantage of the bilinear structure of the problem. For PSO-based method, a search variable is selected among the variable sets causing bilinearity and subjected to particle swarm optimization. For SA-based procedure, a variable neighboring scheme that is similar to a previously used one for the pooling problem is employed. Extensive experiments are conducted to evaluate the performances of these methods and they indicate the success of the proposed solution methods.