The Role of Modeling and Simulation in Controlling Logistics Systems

Charles Sinex, PhD

Johns Hopkins University

Applied Physics Laboratory

11100 Johns Hopkins Road

Laurel, MD 20723-6099


ABSTRACT: Military logistics systems are typically very large and dynamic systems that can have a wide variety of complex interactions. Modeling and simulation tools have been widely used in the design of these systems, as well as to help validate logistics plans as they are developed. However, modeling and simulation tools have not been generally used to help control the actual operation of such logistics systems, although considerable efforts are being devoted to developing sensor suites to monitor system performance and decision planning aids to support rapid re-planning. This paper examines the military logistics system from the perspective of engineering control theory. Control theory is a well-developed field that helps determine the best control methodologies, based on system characteristics, and estimate what performance levels should be achieved. For systems with the characteristics of military logistics systems, model-predictive feed-forward controllers appear to be the best type and the same types of models that would be used to design and test logistics systems would be the types to be used in a model-predictive controller. Complete control of the military logistics system would involve the use of a model-predictive controller linked with sensors to monitor system performance and the re-planning decision aids. This paper discusses how such a model-predictive controller might be implemented and the potential gains it offers to logistics system performance.