A Mixed Discrete PSO Algorithm for Synthesis of Mass Exchange Network with Incompatible Multicomponent

Abstract

A new mixed discrete particle swarm optimization (PSO) algorithm, i.e. DCH-PSO, was presented to solve mixed integer nonlinear programming (MINLP) problems. The main contribution of DCH-PSO is to produce integer speed for discrete variable by randomly choosing it from an integer speed range, which is determined by analysing the velocity updating formula of original PSO, and the influence of historical velocity to the current one was implemented in a probabilistic way. At last, DCH-PSO was applied to synthesize mass exchange network (MEN) involving incompatible multicomponent. The design of orthogonal experiments was applied to study the effect of algorithm’s tuning parameters, and the results show that DCH-PSO is efficient for solving MEN synthesis problems.