Characteristics Of Soft Computing 14383a

Characteristics of Soft Computing

Characteristics • Human expertise • Biologically inspired computing models • New optimization techniques • Numerical computation • New application domains • Model-free learning

• Intensive computation • Fault tolerance • Goal driven characteristics • Real world application

Human expertise • Softcomputing utilized human expertise in the form of fuzzy if then rules as well as in conventional knowledge representations to solve practical problems

Biologically inspired computing models • Biological neural networks-artificial neural networks • Perception, pattern recognition, non linear regression and classification problems

New optimization techniques • • • •

Genetic algorithm Simulated annealing Random search Down hill simplex method

Numerical Computation • Softcomputing mainly rely on numerical computation • Using symbolic techniques in soft computing is an active research area within this field

Model free learning • Neural networks and adaptive fuzzy system have ability to construct models using only target system sample data

Intensive computation • Neuro-fuzzy and soft computing rely heavily on high speed computation to find rules or regularity in data sets • It is a common feature of all areas of computational intelligence

Fault tolerance • Fuzzy and Neural networks exhibit this property • Deletion of a neuron in a neural network or a rule in fuzzy inference system does not destroy the system • Performance quality deteriorates

Goal driven characteristics • Domain specific knowledge helps to reduce the amount of computation and search time • Path leading from current state to the solution does not really matter

Real world applications • Built in uncertainities • Conventional approaches becomes little difficult to be solved • Softcomputing techniques gives satisfactory solutions to real world problems in various disciplines all over the world