Statistical Evaluation of Tactical Supply Chain Planning Methods Based on Scheduling Results

Plant based supply chain planning is often approached hierarchically, decomposed into strategic, tactical and operational planning levels. The detail and amount of information that is incorporated increases as the planning levels become more short-termed. An operational schedule may include additional constraints, such as setup changeovers, inventory holding costs or lost sales penalties, which are not considered at the tactical level. This thesis evaluates two tactical supply chain planning methods. Instead of examining them solely based on traditional performance indicators, such as computational effort or solution gaps, the degree of compatibility with a subsequent scheduling procedure is also considered. The evaluation procedure is such that first, tactical plans are derived using the two methods, which are then passed onto a scheduling program. Finally, performance metrics based on these scheduling results are derived to compare the tactical planning methods.

The computational experiment shows that although one method is clearly superior on a tactical level, much of that superiority is lost after scheduling. It also facilitates the detection of the relationship between specific scheduling parameters and the ranges where they become influential on the scheduling results. This novel evaluation framework provides evidence that implications, that come with the scheduling process, need to be considered when choosing a tactical planning method in a hierarchical supply chain.