A constrained multi-objective particle swarm optimization approach for polystyrene grade transition

The grade transition operation of the polymerization process needs to simultanously satisfy the requirements of safe operation, short transition time, low operating cost, and so on, and therefore it is a multi-objective optimization problem(MOP) with constraints. Usually, a linear weighted sum method is used to transform a multi-objective optimization problem into a single objective problem. In this work, based on the study of the grade transition operation of a continuous styrene polymerization process, two objective functions are constructed focusing on the product quality and the raw material consumption respectively. The constrained multi-objective particle swarm optimization (CMOPSO) approach using control vector parameterization is proposed to solve this multi-objective optimization problem. The process can meet the target quality specification at the end of grade transition operation by adding endpoint constraints on the quality index of the polystyrene. The simulation results confirm that the optimization method can reduce fluctuations of variables, shorten the transition time, reduce the raw materials consumption, and provide multiple grade transition strategies.