Particle Swarm Optimization Algorithm for a Campaign Planning Problem in Process Industries

Abstract

Campaign planning problem (CPP) is to determine the number and length of campaigns for different products over a planning horizon such that the setup and inventory holding costs are minimized. This problem can be found frequently in a multiproduct batch processing plant in the processing industry, such as chemical or pharmaceutical industries. This paper investigates a typical CPP and proposes a hybrid approach of heuristic and particle swarm optimization (PSO) algorithms where the PSO is applied to solve one subproblem with binary variables while the heuristic is applied to the other subproblem with remaining variables by fixing binary variables. As for the evaluation of particles, we take the whole objective function of the primal problem as a fitness function which can be calculated by solving the two subproblems. In implementing the PSO, by designing a “product-to-period” representation for a discrete particle, we redefine the particle position and velocity which are different from the standard PSO. Furthermore, a new strategy is developed to move a particle to the new position. To escape from local minima, a disturbance strategy is also introduced during the iteration process of the PSO. Computational results show that the proposed PSO may find optimal or near optimal solutions for the 180 instances generated randomly within a reasonable computational time.