An efficient hybrid multi-criteria optimization approach for rolling production smoothing of a European food manufacturer

Abstract

This paper introduces an innovative method for multi-objective optimization-based production planning with a rolling horizon for food manufacturing. It features a combination of a heuristic mixed-integer optimization and a metaheuristic optimization. The food industry is characterized by highly fluctuating demand, due to standard and promotion sales volumes, with large periodic variations, resulting in negative impacts on production. State-of-the-art multi-objective solution methods fail to address these complex fluctuations adequately, mainly due to the lack of modelling detail and optimization fine-tuning, i.e. details concerning part-goals and actuating variables. Furthermore, these methods contain static constraints rendering the planning unable to adapt the production system to seasonal (off-) peaks in demand and to consider resource adjustments. In contrast, the presented approach features dynamic capacity-based restrictions and dynamic stock-levels within a given planning horizon. The product volumes are split into sub-dimensions, each with product-specific constraints. In addition to a smoothed production, mid-term stock-levels, setup-costs and the utilization of downstream equipment – evaluated via a deterministic simulation – are optimized simultaneously. The results show a ~40% reduced output variation rate of cost- and labor-intensive key production equipment and a ~430% reduced capacity requirement for downstream equipment.