Performance of an optoelectronic expert system for massively parallel knowledge base applications

Abstract

In this paper a mapped-template logic-based parallel optoelectronic processor is proposed for expert systems in order to achieve high speed and high performance for massively parallel processing of rule-based systems. The proposed system is implemented using two-dimensional space optics symbolic correlator. Also in this paper, an optoelectronic system for a parallel inference engine is presented. The proposed optical inference engine infers conclusions to queries by representing knowledge (facts and rules) in 2D space using a matrix-like formulation. Performance evaluation algorithms for both search and reasoning are presented. Performance enhancement technique is also discussed in this paper. The performance analysis of the proposed system shows its remarkably improved performance compared with multiprocessor electronic systems using conventional programming.