Thapar University 2006 M.E AI & Control Techniques - Question Paper
Thapar Institute of Engineering & Technology
B.Tech Electrical & Instrumentation (1st Year)
Final Term exam
IN004 (AI & Control Techniques)
THAPAR INSTITUTE OF ENGINEERING AND TECHNOLOGY, PATIALA ELECTRICAL & INSTRUMENTATION ENGINEERING DEPARTMENT END SEMESTER EXAMINATION (DECEMBER 13, 2006)
Course Code: IN-004 Course Name: Artificial Intelligence Control
Techniques
Time: 3Hours Marks: 36
Date: December 13, 2006 Instructor: Gagandeep Kaur
Note: Attempt all questions sequentially.
Q1(a) Design Hopfield net for 4bit bipolar patterns. The training patterns are sample S1 = [1 1 -1 -1], S2 = [-1, 1, -1, 1], S3 = [-1,-1,--1,1,1]. Find weight matrix and the energy for three input samples. Determine to which sample S = [-1,1,-1,-1] associates.
(b) What is competitive learning rule? How it is different from Out Star learning? (5,2.5)
Q2(a) Give algorithm for GA. Explain steps in detail with suitable example. Mention engineering application of Genetic Algorithm.
(b) Give steps of Perception Algorithm for several output classes. (5, 2.5)
What do you mean from cardinality of crisp relations? Prove conventional thevenin theorem using fuzzy membership values of current, voltage and resistance. Show all the membership functions graphically assuming their values accordingly. (7)
Q3.
Q4.
The fuzzy sets 1, 2, B3 are shown below. Find their defuzzification patterms using centroid 'method, weighted average method, centre of sums method, centre of largest area. (7)
|
O-v |  |
Q5.(a) Give block diagram of MNN.
(b) Show output net having activation function with threshold factor.
(c) What is the significance of second modification in hetero associative memory neural network?
(d) Give fuzzy equivalence relation for reflexivity and symmetry properties.
(e) First of maxima, method of defuzzification determines smallest value of the domain with maximized membership degree. (True / False)
(1.5,1.5,1.5,1.5,1)
|
Attachment: |
| Earning: Approval pending. |
