Sedigheh Khoshnevis - Department of Computer Engineering, Islamic Azad University, Shahr Qods Branch, Tehran, Iran,
Kavan Sedighiani - Department of Computer Engineering, Shahid Beheshti University, G.C. Tehran, Iran

This paper examines a fuzzy genetic algorithm to construct a fuzzy knowledge-base automatically. This algorithm is a hybrid version of Baldwin and Lamarckian approaches. It includes two steps: genetic evolution and life-time period. Genetic evolution deploys Intuitive Post Condition (IPC) method which offers a chance to construct stronger initial population without losing diversity between initial populations. In this stage, chromosomes will be coded in an effective manner which reduces search space. The life-time period is also employed to improve the rule-base of individuals and guide data-base evolution. In life-time period different types of individuals are designed, any of which has distinct improvement capability that brings some benefits and drawbacks. Experimental results on scheduling data show promising results for this hybrid optimization method.

fuzzy ruled-base systems, genetic algorithm, genetic fuzzy systems, knowledge base