Issue No. 12 - December (2004 vol. 16)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TKDE.2004.95
Albert Mo Kim Cheng , IEEE
We examine the task of constructing bounded-time self-stabilizing rule-based systems that take their input from an external environment. Bounded response-time and self-stabilization are essential for rule-based programs that must be highly fault-tolerant and perform in a real-time environment. We present an approach for solving this problem using the OPS5 programming language as it is one of the most expressive and widely used rule-based programming languages. Bounded response-time of the program is ensured by constructing the state space graph so that the programmer can visualize the control flow of the program execution. Potential infinite firing sequences, if any, should be detected and the involved rules should be revised to ensure bounded termination. Both the input variables and internal variables are made fault-tolerant from corruption caused by transient faults via the introduction of new self-stabilizing rules in the program. Finally, the timing analysis of the self-stabilizing OPS5 program is shown in terms of the number of rule firings and the comparisons performed in the Rete network.
Rule-based systems, knowledge-based systems, expert systems, production systems, real-time, OPS5, self-stabilization, fault tolerance.
S. Fujii and A. M. Cheng, "Self-Stabilizing Real-Time OPS5 Production Systems," in IEEE Transactions on Knowledge & Data Engineering, vol. 16, no. , pp. 1543-1554, 2004.