2110742 Evolutionary Computation 2012

Prabhas Chongstitvatana

Office: Engineering Building 4, floor 18, room 13
phone: 02-2186982
email: prabhas at chula dot ac dot th
class webpage: http://www.cp.eng.chula.ac.th/faculty/pjw/
alternative http://db.tt/iINRD1zS

Class meeting: Thurs 11:00-12:30, Eng Bld 4206, Fri 10:00-11:30pm, meeting room pink Eng Bld 4, Floor 17.

This class discusses evolutionary computation in all aspects. Many established topics in this field will be studied: Genetic algorithms, Genetic programming, Evolution strategies. Advanced topics such as Estimation of distribution algorithms and multiple objective optimisation are included. Theory as well as practical aspects are emphasized.

How the class is conducted?

Because the nature of the topics is the study of algorithms, it renders itself suitable for experimentation. Weekly assignment requires students to run the experiment based on the algorithm in the lecture. It is a graduate class, so students are expected to do a lot of self-study. Individual studies are designed to let students study a specific topic in depth. The results are presented in the group discussion as well as submission of written reports.

Announcement

11 Jan 2013 Project is announced (it is counted as individual study).
23 Feb 2013 Final exam question is here. Due date 1 March 2013, 4pm.

Previous lectures 2010 2008 2006 2005 2003 2000

Weekly lecture

1 Genetic algorithms (Whitley tutorial on GA)    Intro to GA
2 Theoretical basis   GA implementation
3 Probabilistic algorithms
4 Genetic programming read from http://www.genetic-programming.org/    GP in design
5 Evolution strategies slides: one two three four five six
6 Estimation of distribution algorithms (Pelikan slides)
7 EDA 2 Baluja, Population-based incremental learning, CMU CS-94-163, 1994
8 Multiple objective optimisation Deb, NSGA II ieee-trans-ec 2002    slides Multiobjective , Differential Evolution
9-11 Discussion of the recent topic in EC
   Differential Evolution wiki   (paper local copy)
   Particle Swarm Optimization (PSO) wiki
   Ant colony optimization wiki
   No Free Lunch theorem wiki (paper local copy )
   Coevolution ( in nature wiki )   cooperative coevolutionary algorithms ( local copy pdf )

12 Summary

Supplementary materials

1) Georges R. Harik, Fernando G. Lobo, and David E. Goldberg, "The Compact Genetic Algorithm," IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION, VOL. 3, NO. 4, NOVEMBER 1999. (full text)
2) COIN homepage
3) Wattanapornprom, W., Olanviwitchai, P., Chutima, P. and Chongstitvatana, P., "Multi-objective Combinatorial Optimisation with Coincidence Algorithm," IEEE Congress on Evolutionary Computation, Norway, May 18-21, 2009. ( full text )
4) An easier read on COIN description : Chongstitvatana, P., Wattanapornprom, W., Olanviwitchai, P., Sirovetnukul, R., Kampirom N., and Chutima, P., "Coincidence Algorithm for Combinatorial Optimisation and Its Applications," (invited paper), Proc. of Electrical Engineering Conference (33th), Chiangmai, Thailand, 1-3 Dec. 2010. ( full text )
5) A tutorial on multiobjective evolutionary optimization, Zitzler (pdf)
6) Storn, R.; Price, K. (1997). "Differential evolution - a simple and efficient heuristic for global optimization over continuous spaces". Journal of Global Optimization 11: 341–359. (local copy)
7) Wolpert, D.H., Macready, W.G. (1997), "No Free Lunch Theorems for Optimization," IEEE Transactions on Evolutionary Computation 1, 67. ( local copy )
8) R. Paul Wiegand, An Analysis of Cooperative Coevolutionary Algorithms, PhD thesis, Department of Computer Science, George Mason University, 2003. ( local copy pdf )
9) Omidvar, M.N., Xiaodong Li, Zhenyu Yang, Xin Yao, "Cooperative Co-evolution for large scale optimization through more frequent random grouping," IEEE Congress onEvolutionary Computation (CEC), 18-23 July 2010, pp.1-8. ( local copy )
10) Zhenyu Yang, Ke Tang, Xin Yao, "Large scale evolutionary optimization using cooperative coevolution," Information Sciences 178 (2008) 2985–2999. (local copy)
11) Das, S., "Differential Evolution: A Survey of the State-of-the-Art," IEEE Trans on Evolutionary Computation, vol 15, issue 1, Feb. 2011, pp.4-31. (local copy)
12) Manuel Lopez-Ibanez, T. Devi Prasad, Ben Paechter, "Representations and Evolutionary Operators for the Scheduling of Pump Operations in Water Distribution Networks," Evolutionary Computation 19(3): 429–467. (local copy)

Homework

1 Try your hands on solving TSP with GA. Go to this site. It is Java (applet) based. http://www.ads.tuwien.ac.at/raidl/tspga/TSPGA.html
2 Find out what kind GA recombination operator work for TSP problems?
3 Implement and run compact GA on 6xtrap-5 (30 bits) problem. Discuss divergence, parameters setting issues
4 Read NSGA II, prepare to discuss this algorithm.
5 Read No Free Lunch paper (the first few pages) and prepare to discuss how the authors proof their theory.
6 Read Cooperative Coevolutionary Algorithm, prepare to discuss how a normal EC be changed into a Coevolve one.
7 Prepare a presentation of applications of EC for next week class.

Project

Try using EC or inventing (or mixing up) algorithms of your own, to solve the flowshop problems. Using this reference to compare your method with. You don't have to implement your own program. Use any available method to do it. Warning, the benchmarks are tough to beat. I will give more details about which data set we will use later. Do your best. You must write around 4-5 pages of English report about your experiment. Here is the reference:
Quan-Ke Pan, Ruben Ruiz, "Local search methods for the flowshop scheduling problem with flowtime minimization," European Journal of Operational Research, 222 (2012) 31-43. ( pdf )

External references

Applet (animation) for GA

http://www.rennard.org/alife/english/gavgb.html

Genetic Algorithm Viewer 1.0 is a demonstration applet of the functioning of a Genetic Algorithm (GA). It aims at showing the power of GA and of the main mechanisms used while permitting a certain form of visualization of the general functioning.
http://www.ads.tuwien.ac.at/raidl/tspga/TSPGA.html
Visualisation of Genetic Algorithms for the Traveling Salesman Problem in Java
http://www.stellaralchemy.com/ice/
The applet executes what is called a genetic algorithm (GA). To facilitate understanding, every step in the GA is animated. This particular GA evolves solutions to a simple board game. This program also detects and reports if any irreducibly complex (IC) solutions arise during the evolution of its population of solutions to the board game.
http://www.obitko.com/tutorials/genetic-algorithms/example-function-minimum.php
Minimum of Function: The problem is expressed as looking for extreme of a function. Some function is given and GA tries to find minimum of the function. For other problems we just have to define search space and the fitness function which means to define the function, which we want to find extreme for.

Source code for Differential Evolution (from Storn) http://www1.icsi.berkeley.edu/~storn/code.html

Assessment

homework 20
individual study 30
final exam 50

Text

1. Mitchell, M., An introduction to genetic algorithms, MIT press, 1996.
2. Goldberg, D., The design of innovation, Kluwer pub, 2002.
3. Holland, J. H., Adaptation in natural and artificial systems, MIT press, 1992.
4. Koza, J., "Genetic Programming Vol 1, 2, 3", MIT Press, 1992, 1994, 1999.
5. Goldberg, D., Genetic algorithms, Addison-Wesley, 1989.
6. Banzhaf, W., Nordin, P., Keller, R. and Francone, F., Genetic Programming: An Introduction, Morgan Kaufmann, 1998.
6. Winter, G., Periaux, J., Galan, M., Cuesta, P. (eds), Genetic algorithms in engineering and computer science, John Wiley, 1995.

Uptodate handouts on various current researches will be distributed in the class.

Tools

galib247.zip
myga example
I have mentioned Heuristic Lab as a good tool. Try it.

last update 23 Feb 2013