Soft Comp is a combination of Evolutionary computing, Neural Network and Fuzzy logic. .

AI magic mathematical NN structure which has ability to learn and self correct using any algorithm. The AI also part of Soft Comp , bz both have NN,

Refer Figure no.1

Article APPLICATIONS OF SOFT COMPUTING IN VARIOUS AREAS

*Soft computing* is concerned with the use of approximate calculations to provide imprecise but workable solutions to problems which are too complex to be exactly solved in a reasonable amount of time.

*Artificial Intelligence*, on the other hand, deals with the development of systems which able to perform tasks normally requiring human intelligence.

Since some of the tasks undertaken by AI are too complex to be solved exactly, SC may be used to solve these problems.

SC can therefore be seen as overlapping with a subset of the problems arising in the field of AI.

Dear Shounak Datta ,

Thank you fot your response. However, why FLC, ANN, Genetic algorithms Machine Learning and Swarm Optimization are a subset in SC and AI ?

Dear Saad,

The methods you have mentioned are all used to accomplish AI tasks such as classification, clustering, etc. Hence, these fall within the field of AI. Each of these methods are also SC methods for different reasons:

FLC - uses fuzzy logic and fuzzy rules instead of crisp logic.

ANN, GA, SO - only finds locally optimal solutions.

Machine Learning is more of a subfield of AI and not an SC technique.

Hope this helps.

Dear Shounak Datta ,

According to Wikipedia, The principal constituents of Soft Computing (SC) are Fuzzy Logic (FL), Evolutionary Computation (EC), Machine Learning (ML) and Probabilistic Reasoning (PR), with the latter subsuming belief networks and parts of learning theory.

Please, refer to :

https://en.wikipedia.org/wiki/Soft_computing

We can view AI as the base set out of which subsets are generated i.e. evolutionary, NN , Nature inspired, soft computing, ML etc.

Artificial intelligence is a broad term which primarily focus on making use of human learning patterns to solve complex computing problems by appropriate modelling method. If modelling method is Nature inspired then we can categories such AI methodologies into a area call soft computing. Soft computing when applied to complex problems promise near to optimal acceptable solutions as done by many living creatures in nature which by and large use basic understanding to achieve some solution to their problems rather than stucking on for some particular global optimal solution.