An Introduction to Soft Computing Techniques
A new approach called soft computing was recently developed to perform the process of computation. This technique was designed through artificial, as well natural concepts. The various approaches that come under the soft computing approach are Probabilistic Reasoning, Machine Learning, Evolutionary Computation, Support Vector Machines, Neural Network, and Fuzzy Logic. There are several advantages to soft computing, such as good tolerance of imprecision and uncertainty, cognitive ability, and strong learning. For this reason, this technology appears to be an emerging technology.
Traditional computing is hard computing, but there are certain limitations in hard computing such as approximation, partial truth, uncertainty, and imprecision. The concept of soft computing was developed to overcome these limitations. The model of soft computing was developed by mimicking the human mind.
Researcher Dr. Loth Zadeh (1965), developed this concept of soft computing. The technique is widely adopted in multidisciplinary fields. The main objective for developing this technique is to modify Artificial Intelligence to obtain a newer approach called Computational Intelligence. Basically, the technique of soft computing originated in 1981. Dr. Zadeh explained soft data analysis in his first work (Zadeh, 1997). After this initial description the technique of soft computing has since progressed. Through merging several fields, such as Probabilistic Computing, Genetic and Evolutionary Computing, Neuro-Computing, and Fuzzy Logic into a single multidisciplinary approach, the concept of soft computing evolved. This is the definition given for Soft Computing by Dr. Zadeh. The ultimate aim for developing this soft computing was to generate intellectual machines to solve mathematical and nonlinear problems (Zadeh, 1993, 1996, 1997).
There are two major advantages of the soft computing technique. Nonlinear problems, for which there is a lack of mathematical model, can be resolved by means of this technique. The basic knowledge that is found in human-like learning, understanding, recognition, and cognition were employed in this computing. Based on this information the generation of intellectual systems is achieved. Automated designed systems and self-tuning systems are some of the intelligent systems. The technique of soft computing is one of the newer concepts in the field of science. It has attained tremendous growth and advancement which is beyond what Dr. Zadeh, the author who initiated this concept, thought possible. For instance, fields such as Multi-Valued Logic, Evidential Reasoning, Probabilistic Computing, Evolutionary Computing, Neural Networks, Rough Sets, Fuzzy Sets, etc., which must be included in this soft computing technique, have been explained in certain literature (Kacpzyk, 2001). Within the normal soft computing concept, the introduction of Immune Network Theory and Chaos Computing was to develop a newer concept, referred to as Extended Soft Computing (ESC). The concept of ESC was developed by (Dote et al. 2000). In the previous section, soft computing and its developers were explained. The technique of ESC was developed to overcome reactive and cognitive AI, as well as to solve complex systems. Traditional probability computing techniques process soft systems computations.
The fundamental requirement for developing this soft computing was fulfilled by Fuzzy Logic. The advancement in Fuzzy Logic led to the development of Type-2 Fuzzy Logic (Mendel, 2001). Recently, the emergence of a newer science, based on Biotic and Bios Systems, is trending. The idea of Bios Computing plays a significant role in the technique of soft computing. It has been proved through many authors’ expertise, who clearly explained the reason for replacing the usual computation with soft computing, and how it has been achieved.