Lecture #9: Fuzzy Logic & Neural Nets s1

Total Page:16

File Type:pdf, Size:1020Kb

Lecture #9: Fuzzy Logic & Neural Nets s1

Monday February 10, 2004

DSES-4810-01 Intro to COMPUTATIONAL INTELLIGENCE & SOFT COMPUTING

Instructor: Prof. Mark J. Embrechts (x 4009 or 371-4562) ([email protected]) Office Hours: Thursday 10-11 am (CII5217), or by appointment. Class Time: Monday/Thursday: 8:30-9:50 (Amos Eaton 216) TEXT: J. S. Jang, C. T. Sun, E. Mizutani, “Neuro-Fuzzy and Soft Computing,” Prentice Hall, 1996. (1998) ISBN 0-13-261066-3

LECTURES 8&9: GENETIC ALGORITHM APPLICATIONS

This lecture will illustrate several applications of genetic algorithms: 1. Hillis search function 2. Traveling salesman Problem 3. Gene Alignment Problem 4. Peaks Function 5. Michalewicz library of difficult problems 6. Finding Cellular Automata with GAs

We also point out what it takes to make GAs work for difficult optimization problems: 1. Apply the right encoding of the problem in order to avoid penalty functions 2. Use floating point GAs 3. Use the right and specialized genetic operators for your problem 4. Use a smart selection procedure 5. Consider the cases where evolutionary algorithms outperform GAs 6. Use immuno-computing to avoid penalty functions

These lectures will also point out the differences between genetic algorithms, evolutionary algorithms, and evolutionary programming. Nowadays these three related fields are lumped under evolutionary computing.

Quiz (February 19)

A 20-minute quiz will be polled based on Chapter 2 of the textbook (a 1 page cribsheet, with anything you can fit on both sides is allowed).

H omework #3: (February 23)

Write the objective function for the peaks function in [-3,3], and search for the maximum and the minimum with a genetic algorithm. Most of the work is already coded up for you in the program that is attached to the lectures for this week on www.rpi.edu/~embrem. The peaks function is described on p. 179 of the text (and repeated here for your convenience). A MATLAB script displaying the peaks function is also posted on the web site. The C program that is posted on the CI website is significantly edited for readability from a code posted on the web (ftp.uncc.edu, directory coe/evol, file prog. C). The code should run on most standard compilers (you possibly have to remove the #include directive on UNIX machines).

1 PEAKS FUNCTION:

2 2 2  x  2 2 1 2 2 x  f x, y  31 x e x  y1 10  x 3  y 5 e x  y  ex1  y  5  3

Handout:

1. David Fogel,” What is Evolutionary Computing?” IEEE Spectrum, pp. 26-32, February 2000.

Deadlines

January 22 Homework Problem #0 (web browsing) January 29 Project Proposal February 16 Homework Problem #1 (paper) February 19 Quiz #1 (20 minutes, Chapter 2 of Book) February 23 Homework Problem #2 (peaks function)

2

Recommended publications