DISPO 4.0 | Simulation-Based Optimization of Stochastic Demand Calculation in Consumption-Based Material Planning in the Capital Goods Industry

Abstract

This paper presents a digital material planning approach, utilizing simulation-based optimization to select and parametrize article specific demand forecasting methods. Demand forecasts are the basis of material requirements planning in consumption-based material planning, and are an essential lever for efficient inventory and order calculation. Despite their acknowledged potential, digital tools for optimized demand calculation are still lacking in practice. Thus, the goal of the presented approach to provide an application-oriented method to optimally select and parametrize state-of-the-art forecasting methods, based on product-specific demand data. In this approach, a rule-based selection heuristic is combined with static simulation of demand time-series and a metaheuristics-based optimization of forecasting parameters, to provide automatically optimized article-specific demand forecasts. Case studies for two companies in the capital goods industry evaluate and quantify the application potential. The results point to significantly improved, item-specific demand planning.

Keywords

• demand planning
• exponential smoothing
• forecasting
• parameter optimization
• simulation