Considers the application of genetic algorithms to the automaticgeneration of fuzzy process controllers. In contrast to priorgenetic-fuzzy systems which require that every input combination beenumerated, we employ a novel encoding scheme which maintains only thoserules necessary to control the target system. The key to this method isto represent each fuzzy system as an unordered list of an arbitrarynumber of rules. This compact rule base methodology, introduced byCooper and Vidal (1993) to generate a controller for the cart andsingle-pole problem, is applied to the more complex cart andjointed-pole problem. By using a compact rule base representation,successful systems evolve quickly, increasing the likelihood of successwhen applied to complex problems
展开▼