Dynamic Task Assignment in Distributed Manufacturing Systems

Abstract

Paradigms of decentralized and “on-line” control developed in the context of distributed computing systems will be relevant to distributed manufacturing systems of the future. In this paper, we explore certain aspects of decentralized task allocation in distributed systems using formal mathematical models and analyses to justify effective control strategies. We analyze two schemata for task allocation in distributed manufacturing systems modeled as stochastic processes. The first schema has to do with the selection of on-line policies for task assignment in distributed environments with the feature that the uncertainty in the status of machines can be removed by incurring query costs. In the second schema, optimal task allocation in the distributed batch processing model of the manufacturing system is resolved using techniques from Markov decision theory.