The Fuzzy Inference Engine Developed Using Genetic Algorithms and Expert Knowledge

In this work realization of fuzzy inference system (FIS) with the use of genetic algorithms and expert knowledge is presented. Every day many IT companies are facing problems that are characterized by uncertainties. Often decisions of these problems are drawn by using subjective judgements that may lead to unsuccessful result of entire project. Software implementation of FIS is proposed to overcome these difficulties and to give a proper account to uncertanties. On top of this, genetic algorithms help to fit FIS through learning process to a specificity of current problem. Moreover, expert knowledge is used to reduce influence of uncertainties that may still exist once the learning process is finished. In this report different approaches of FIS improvement by genetic algorithms and possibility of gathering initial data from file by using fuzzy clustering algorithm are studied. Usage of web-technologies helps to ensure easiness of FIS realization for further utilization by experts and engineers, who maintain the operability and actuality of the system. This report contains 64 pages, 3 chapters, 30 illustrations, 6 tables, 19 references and 9 appendices. Keywords: fuzzy logic, fuzzy inference systems, genetic algorithms, expert knowledge, fuzzy clustering.