Automatic synthesis of distributed transition systems

Abstract

This thesis investigates the synthesis problem for two classes of distributed transition systems: synchronous products and asynchronous automata. The underlying structure of these models consist of local automata synchronizing on common actions. The synthesis problem discussed is as follows: Given a global specification as a transition system TS and a distribution pattern ∆, find a distributed transition system over ∆ whose global state space is ‘equivalent’ to TS . As criteria for the correctness of the (distributed) implementation vs. the specification (i.e., their ‘equivalence’) we use: transition system isomorphism, language equivalence, and bisimilarity respectively. In particular, the synthesis of asynchronous automata modulo language equivalence is a notoriously hard problem solved by Zielonka at the end of the 80s. One of the motivations behind our work was to bring this theory closer to practical applications. From the theoretical point of view, we conduct a detailed analysis of the synthesis problem for both models of distributed systems, look at effective algorithmic approaches and draw a map of computational complexity results. E.g., we provide several matching lower and upper complexity bounds for the distributed implementability problem. From the practical perspective, we provide prototype implementations for most of the synthesis algorithms discussed in the thesis. Moreover, we offer assistance when a given specification is not distributable by trying to modify this specification such that distributed synthesis can be applied. By using several heuristics to overcome the classical state space explosion, we are able to automatically generate small distributed algorithms for problems such as mutual exclusion.

Extracted Key Phrases

19 Figures and Tables

Cite this paper

@inproceedings{Stefanescu2006AutomaticSO, title={Automatic synthesis of distributed transition systems}, author={Alin Stefanescu}, year={2006} }