Swarm Synthesis of Convergence for Symmetric Protocols

This paper presents a novel non-deterministic method for algorithmic addition of convergence to non-stabilizing symmetric protocols. The proposed method exploits randomization and parallelization in order to expand the scope of the search for self-stabilizing versions of non-stabilizing protocols. Such a non-deterministic method enables an embarrassingly parallel framework that exploits the computational resources of computer clusters for automated design of self-stabilizing protocols. We have implemented our approach in a software tool and have synthesized several new self-stabilizing solutions for well-known protocols in the literature (e.g., maximal matching, graph coloring and leader election on a ring). Our case studies demonstrate that the proposed method is able to automatically generate self-stabilizing versions of non-stabilizing protocols in cases where existing automated methods fail. As a result, the proposed approach increases the likelihood of success in synthesizing the self-stabilizing versions of non-stabilizing protocols.