Application of Genetic Algorithm and Rough Set Theory for Knowledge Extraction

This paper proposes a hybrid approach using the rough set theory and genetic algorithm (RS-GA) for knowledge extraction as one part of a substation level decision support system. The technique involved a process which learns and extracts knowledge from a set of events into a form of rules to identify the most probable faulted section in a network. Numerous case studies performed on a simulated distribution network [1] that consists of several relays models [2] using PSCAD/EMTDC have revealed the usefulness of the proposed technique for fault diagnosis. The test results demonstrated that the extracted rules are capable of identifying and isolating the faulted section and hence improve the outage response time.