A Fast Reacting, and Distributed Architecture for Supply Chain Management

In this proposal, we give a detailed account of three integrated components of our proposed solutions. The first component, dynamic lateral networks, introduces a framework for dynamic allocation of resources among nodes in an MRO domain. This framework allows fast reaction to dynamically changing environments. We will define the “choice parameter”; the choice parameter reflects, in statistical terms, the role of human agents in resource allocation and addresses the need for quick adaptation to changing requirements. Next, we propose a randomized parallel architecture based on probabilistic Petri Nets which captures spatial and temporal distributed characteristics of supply chain management problems. They will be the basis for a new generation of discrete event simulation and supply chain management software tools. Randomizing Petri Nets in such a way that permits parameterization from trace data is a promising approach; this will lead to more realistic models for MRO applications. The third component is the robustification of policies under model uncertainties and constrained capacities; we will develop robust policies under limited capacity, which is the case in most practical applications.