Ant colony algorithm for an industrial batch sizing and sequencing problem in injection molding system

This study addresses a real-life problem involving an injection machine capable of producing various product types, each requiring a unique configuration. The challenge is to organize batch production to meet demand, maintain buffer levels within specified ranges, and maximize resource efficiency. The model accounts for sequence-dependent setup times and stock coverage requirements. The main objectives are to reduce backorders, minimize stockouts, and maximize machine utilization. Three versions of Ant Colony Optimization algorithm (ACO) are proposed to determine the optimal sequence of machine states throughout the time horizon. The machine can operate in one of three states: production, setup, or idle. The algorithms are validated through experiments on instances inspired by real-world scenarios. To evaluate the performance of the ACO, the results obtained from the testing phase are compared with those of the ILP mathematical model, both in terms of solution quality, based on a multi-objective function, and the computational time required. The study concludes with recommendations for future work, including extending the model to consider parallel machines that share common setup resources. The paper concludes with an analysis of the numerical experiments and suggestions for future research directions.