A particle swarm optimization for minimizing total earliness/tardiness costs of two-stage assembly flowshop scheduling problem in a batched delivery system

This paper considers a two-stage assembly flowshop scheduling problem in which a number of single-item is ordered with a due date. Each of them is manufactured by assembling several different pieces that must be processed on first stage in a flowshop scheduling system. When all parts of each product are completed in the first stage, they are assembled into a final product on an assembly machine in the second stage. Due to delivery cost, in order to reduce this cost, completed products can be held until completion of some other products to be delivered in a same batch. The proposed problem addresses scheduling a set of operation with specific due date that must be processed on all of the machines in the two stages and dispatched to customers in a batch delivery system. The aim is to minimize total weighted earliness/tardiness and delivery costs. As the problem is demonstrated to be NP-hard, a genetic algorithm (GA) and a particle swarm optimization (PSO) are presented to solve the problem in real scales. In order to evaluate the proposed algorithms, a number of problems are solved by them and the results are compared. The computational results illustrate that the proposed PSO has a qualifier performance than the GA.