Effects of forecast errors on optimal utilisation in aggregate production planning with stochastic customer demand

TL;DR

This paper investigates how forecast errors impact the optimal utilization in aggregate production planning under stochastic demand, highlighting the importance of adjusting planned utilization to reduce costs in complex production systems.

Contribution

It evaluates the effect of forecast errors on the optimal utilization factor in stochastic demand environments and discusses how adjusting this factor can improve overall costs.

Findings

Forecast errors significantly affect optimal costs.

Adjusting planned utilization reduces costs under stochastic demand.

Negative impacts of forecast errors vary with system structure.

Abstract

The hierarchical structure of production planning has the advantage of assigning different decision variables to their respective time horizons and therefore ensures their manageability. However, the restrictive structure of this top-down approach implying that upper level decisions are the constraints for lower level decisions also has its shortcomings. One problem that occurs is that deterministic mixed integer decision problems are often used for long-term planning, but the real production system faces a set of stochastic influences. Therefore, a planned utilisation factor has to be included into this deterministic aggregate planning problem. In practice, this decision is often based on past data and not consciously taken. In this paper, the effect of long-term forecast error on the optimal planned utilisation factor is evaluated for a production system facing stochastic demand and the benefit of exploiting this decision's potential is discussed. Overall costs including capacity, backorder and inventory costs, are determined with simulation for different multi-stage and multi-item production system structures. The results show that the planned utilisation factor used in the aggregate planning problem has a high influence on optimal costs. Additionally, the negative effect of forecast errors is evaluated and discussed in detail for different production system environments.