Parallel model-based diagnosis using PVM
The present paper outlines the PVM implementation of a particular approach to model-based diagnosis which uses a Petri net model of the system to be diagnosed. Parallel backward reachability analysis on the state space of the net is used to explain the misbehavior of the modeled system. The analysis algorithm is based on the automatic identification of the parallelism present in the structure of the Petri net model. Starting from the above information, an MIMD message passing program is automatically constructed. PVM has proved to be a useful tool for implementing the above parallel programs; in particular, we tested the approach by implementing a parallel program for a car fault diagnosis domain.
Unable to display preview. Download preview PDF.
- 1.SCOTCH Home Page: http://www.labri.u-bordeaux.fr/∼ pelegrin/scotch.Google Scholar
- 2.C. Anglano and L. Portinale. B-W analysis: a backward reachability analysis for diagnostic problem solving suitable to parallel implementation. In Proc. 15th Int. Conf. on Application and Theory of Petri Nets, LNCS 815, pages 39–58, Zaragoza, 1994. Springer Verlag.Google Scholar
- 3.W. Hamscher, L. Console, and J. de Kleer. Readings in Model-Based Diagnosis. Morgan Kaufmann, 1992.Google Scholar
- 4.T. Murata. Petri nets: Properties, analysis and applications. Proceedings of the IEEE, 77(4):541–580, 1989.Google Scholar
- 5.L. Portinale. Petri net models for diagnostic knowledge representation and reasoning. PhD Thesis, Dip. Informatica, Universita' di Torino, 1993. Available as anonymous ftp at ftp.di.unito.it in /pub/portinal.Google Scholar
- 6.L. Portinale. Petri net reachability analysis meets model-based diagnostic problem solving. In Proc. IEEE Intern. Conference on Systems, Man and Cybernetics, pages 2712–2717, Vancouver, BC, 1995. Extended version to appear on IEEE Trans. on SMC.Google Scholar