Deadlock-free distributed control for repetitive flows

Abstract

This paper addresses resource allocation issues that are at the heart of a methodology proposed for the design of distributed control of a class of repetitive processes. In particular, conditions, sufficient for deadlock-free and starvation-free execution of a system of closed, sequential,. repetitive processes are considered. The relationship between system resource capacities and an initial state as well as priority rules allocation is investigated. Finally, following developed conditions sufficient for design and allocation of priority rules a distributed buffer control policy that restricts a process entering a buffer in a specific sequence is proposed.