A stochastic production planning optimization model for studying the value of flexibility under demand uncertainty

This study proposes a mixed-integer linear programming for studying the benefits of flexibility in a production system under demand uncertainty. The model is representative of consumer package goods industries. A rolling horizon approach is used to simulate the planning process. A 2^k factorial design and a factorial regression are performed to analyze the impact on profit due to changes in levels of flexibility and uncertainty factors. The evaluated factors include setup times, production capacity, replanning frequency, number of production lines, and the process time; demand estimate bias and standard deviation are the included uncertainty factors. This study was applied to a real snack food company as a case study. The results show that bias and replanning frequency are the factors that most impact profit, and replanning more frequently is the most significant strategy to reduce the negative impact on profit due to forecasts with high bias.