Deadlock avoidance policies for automated manufacturing cells

The problem of deadlock resolution arising in manufacturing environments organized according to the typical process-layout has only recently been undertaken by the scientific community. Furthermore, deadlock avoidance methodologies, which seem to be the most appropriate deadlock-handling strategies in this particular context, unfortunately, in the general case, suffer from increased computational complexity results in heuristic solutions and/or reduced performance. This paper proposes an analytical framework for designing scalable and yet provably correct deadlock avoidance policies for a subclass of resource allocation systems (RAS), which is appropriate for the study of the deadlock problem in the context of the automated manufacturing cell (AMC). The framework is used to prove the correctness of the resource upstream neighborhood (RUN) policy, developed in the authors' research program.