Agent-based modelling in capacitated lot sizing problem with sequence dependent setup time

Abstract

Setups are indispensable in production, but consume substantial amount of productive time. It is vital to consider sequence dependent setup times for determining production lots to satisfy demand for diversified product types on parallel production lines. Optimisation is the most used technique to generate detailed schedules for such sequencing problems. But the feasibility of the solution is not guaranteed under uncertainty conditions. Thus, evaluating the solution configurations under uncertainty is necessary to confirm feasibility. In this paper, an agent-based, discrete event simulation technique is used to develop a flow-shop model which faces demand for multiple product varieties, and has sequence dependent setup time. The simulation aims at evaluating various results in the solution space of an optimisation model to check their feasibility under various uncertain conditions in the form of setup time, processing time, and demand. Based on the results obtained from this optimisation-simulation model, it was observed that uncertainty influences cost parameters such as overtime cost, holding cost, and lost sales.