Soft Computing

Introduction to Soft Computing

Prof. Debasis Samanta

Department of Computer Science & Engineering
IIT Kharagpur

1
INTRODUCTION TO SOFT COMPUTING
• Concept of computaton
• Hard computng
• Sof computng
• How sof computng?
• Hard computng vs. Sof computng
• Hybrid computng

Debasis Samanta
CSE
2
IIT Kharagpur
CONCEPT OF COMPUTATION
 
Antecedent Computng Consequent

Control Acton

Figure: Basic of computng
 
 is a mapping functon.
 is also called a formal method or an algorithm to solve a problem.

Debasis Samanta
CSE
3
IIT Kharagpur
Important characteristics of computing

• Should provide precise soluton.

• Control acton should ne unambiguous and accurate.
• Suitable for problem, which is easy to model mathematcally.

Debasis Samanta
CSE
4
IIT Kharagpur
Hard computing

• In 1996, L. A. Zade (LAZ) introduced the term hard computng.

• According to LAZ: We term a computng as Hard computng, if
 Precise result is guaranteed.
 Control acton is unambiguous.
 Control acton is formally defned (i.e., with mathematcal model or

algorithm).

Debasis Samanta
CSE
5
IIT Kharagpur
Examples of hard computing
• Solving numerical problems (e.g., roots of polynomials,
integraton, etc.).
• Searching and sortng techniques.
• Solving computatonal geometry problems (e.g., shortest
tour in a graph, fnding closet pair of points given a set of
points, etc.).

• many more…
Debasis Samanta
CSE
6
IIT Kharagpur
Soft computing
• The term sof computng was proposed by the inventor of
fuzzy logic, Loti A. Zadeh. He describes it as follows.
Defnition t: Soft computing
Sof  computng  is  a  collecton  of  methodologies  that  aim  to  exploit  the 
tolerance for imprecision  and uncertainty to achieve tractability, robustness, 
and  low  soluton  cost.  Its  principal  consttuents  are  fuzzy  logic,  neuro-
computng, and probabilistc reasoning. The role model for sof computng is 
the human mind.

Debasis Samanta
CSE
7
IIT Kharagpur
Characteristics of soft computing
• It does not require any mathematcal modeling of problem solving.
• It may not yield the precise soluton.

• Algorithms are adaptve (i.e., it can adjust to the change of dynamic

environment).
• Use some biological inspired methodologies such as genetcs, evoluton,
Ant’s behaviors, partcles swarming, human nervous system, etc.).

Debasis Samanta
CSE
8
IIT Kharagpur
Examples of soft computing

Sof 
⋮
 
computng

Example: Hand writen character recogniton 
(Artfcial Neural Networks)

Debasis Samanta
CSE
9
IIT Kharagpur
Examples of soft computing

Sof  Bank with 
computng maximum return
⋮
 

Example: Money allocaton problem 
(Evolutonary Computng)
Debasis Samanta
CSE
10
IIT Kharagpur
Examples of soft computing

Example: Robot movement 
(Fuzzy Logic)

Debasis Samanta
CSE
11
IIT Kharagpur
How soft computing?
• How a student learns from his teacher?
• Teacher asks questons and tell the answers then.
• Teacher puts questons and hints answers and asks whether the
answers are correct or not.
• Student thus learn a topic and store in his memory.
• Based on the knowledge he solves new problems.
• This is the way how human brain works.
• Based on this concept Artfcial Neural Network is used to solve problems.
Debasis Samanta
CSE
12
IIT Kharagpur
How soft computing?
• How world selects the best?
• It starts with a populaton (random).
• Reproduces another populaton (next generaton).
• Rank the populaton and selects the superior individuals.
• Genetc algorithm is based on this natural phenomena.
• Populaton is synonymous to solutons.
• Selecton of superior soluton is synonymous to exploring the optmal
soluton.
Debasis Samanta
CSE
13
IIT Kharagpur
How soft computing?
• How a doctor treats his patent?
• Doctor asks the patent about sufering.
• Doctor fnd the symptoms of diseases.
• Doctor prescribed tests and medicines.
• This is exactly the way Fuzzy Logic works.
• Symptoms are correlated with diseases with uncertainty .
• Doctor prescribes tests/medicines fuzzily.

Debasis Samanta
CSE
14
IIT Kharagpur
Hard computing vs. Soft computing
Hard computng Sof computng
 It  requires  a  precisely  stated   It  is  tolerant  of  imprecision, 
analytcal  model  and  ofen  a  lot  of  uncertainty,  partal  truth,  and 
computaton tme. approximaton. 

 It  is  based  on  binary  logic,  crisp   It  is   based  on  fuzzy  logic,  neural 
systems,  numerical  analysis  and  nets and probabilistc reasoning.
crisp sofware.
 It has the characteristcs of  It has the characteristcs of
precision and categoricity. approximaton and dispositonality.

Debasis Samanta
CSE
15
IIT Kharagpur
Hard computing vs. Soft computing

Hard computng Sof computng

 It  is deterministc.  It incorporates stochastcity. 
 It  can  deal  with  ambiguous  and 
 It requires exact input data. noisy data.
 It is strictly sequental.  It  allows parallel computatons.
 It produces precise answers.  It can yield approximate answers

Debasis Samanta
CSE
16
IIT Kharagpur
Hybrid computing
• It is a combinaton of the conventonal hard computng and 
emerging sof computng.

HC SC

Hybrid computng

Figure: Concept of Hybrid Computng

Debasis Samanta
CSE
17
IIT Kharagpur
In this course…
• You will be able to learn
• Basic concepts of Fuzzy algebra and then how to solve problems using
Fuzzy logic.

• The framework of Genetc algorithm and solving varietes of

optmizaton problems.

• How to build an artfcial neural network and train it with input data to
solve a number of problems, which are not possible to solve with hard
computng.

Debasis Samanta
CSE
18
IIT Kharagpur
Thank You!!

Debasis Samanta
CSE
19
IIT Kharagpur