A Horizon Sensitivity Simulation of a Multistage Linear Programming Aggregate Operations Model

Publisher Website
Google Scholar
Abstract
A firm that uses an aggregate operations planning model faces an important problem. Because it is multiperiod, solutions of a model with different horizon lengths can result in significantly different decisions, especially if aggregate shortages are allowed. This study illustrates methods of reducing the horizon length effects of a general linear programming model used by a plastic product manufacturer. As shown, that model originally had solutions of dubious value because of horizon effects. These effects were greatly reduced by eliminating aggregate shortages, but that precluded their use during peak seasonal demands. A more effective method was to penalize shortages during only the last period of a horizon. This improved the model's solutions without increasing its complexity, and also made it acceptable to responsible managers. Similar benefits should occur to users of analogous operations models which allow aggregate inventory shortages.

This publication has 0 references indexed in Scilit: