Distributed supply chain simulation using a generic job running framework

Abstract

For supply chain performance simulation that involves aggregating results from multiple runs of the same underlying model, simulation iterations can be distributed to networked computing resources to achieve significant speedup. This paper presents a generic distributed job running framework that facilitates such high performance supply chain simulation. We first introduce a supply chain modeling and simulation tool developed by IBM Research, and summarize the strategy to enhance it. A closer look is then taken at a generic job running framework we designed and how it was used to bring the distributed simulation capability to the tool. After reviewing an ongoing effort to integrate the new tool with the IBM MathGrid environment, we conclude the paper with a brief discussion of our future work.