Large deviations analysis for distributed algorithms in an ergodic Markovian environment (Q843968)

This paper deals with a large deviations analysis of deadlock phenomena occurring in distributed systems sharing common resources. In this model transition probabilities of resource allocation and deallocation are time and space dependent. The organization of this paper is as follows: the authors discuss probabilistic model in Sect. 2. In Sect. 3 they prove a Large Deviations Principle. Deadlock phenomenon analysis is done rigorously with much details in Sect. 4. Then they illustrate with the two-stacks model and show an example where the system has a stable attractor which is a limit cycle.