COMPARATIVE STUDY ON APPLICATION OF DIFFERENTIAL EVOLUTION AND PARTICLE SWARM OPTIMIZATION FOR TUNING CONTROLLERS OF BALL AND BEAM I. SYSTEM AND ROBOT MANIPULATOR

Abstract

Soft Computing is a collection of techniques covering many areas that come under different classifications in computational intelligence. The computing techniques belong to the various fields such as computer science, machine learning and engineering fields, which undergoes study of model and their analysis of associated complex phenomena. Such benefits have not come forward through the conventional methods. They are far behind in yielding low cost complete solutions. Soft computing techniques use soft techniques contrasting it with classical artificial intelligence of its counterpart. The techniques are developed on the information processing in biological systems. The tasks like surrounding recognition, act according to the plans as per the ideas thought of in order to survive is the eloquent feature of the complex biological information system in humans.. The information processing involves both logical and intuitive processing. Logical processing is what conventional computers are good at, but they are far behind in capability for the later as compared to human beings. The three features are required for any computing system to have human like information processing capability: openness, robustness and real time processing[1]. Openness of a system is its capability to cope with circumstantial and random changes encountered in the real world and also allowing + it to extend on its own. If a system has tolerance and also remains stable even if it meets with segregated, incomplete or imprecise information then it is said to be as robust. A system has real time processing characteristic if it reacts in a considerable amount of time if encountered with an event. Real world computing (RWC) systems are said to have these three features. A RWC system is therefore capable of representing the information in distributed manner, processing parallel in huge amount when required, adapting in order to organize itself and learning at the same time to achieve flexibility in information processing. Thus, RWC systems incorporate the soft computing techniques as key ingredient