Presenter : Tsung-Yu Ho 2009.12.04. Review Mixing Problem Categories Crossover as a Mixer (√) Crossover as a Innovator (√) Crossover as a Disrupter All.

Slides:



Advertisements
Similar presentations
Genetic Algorithm.
Advertisements

A First Course in Genetic Algorithms
Genetic Algorithm with Limited Convergence 1 Simple Selectorecombinative GAs Scale poorely on hard problems (multimodal, deceptive, high degree of subsolution.
Genetic Algorithms Representation of Candidate Solutions GAs on primarily two types of representations: –Binary-Coded –Real-Coded Binary-Coded GAs must.
Genetic Algorithms as a Tool for General Optimization Angel Kuri 2001.
Computer Science Genetic Algorithms10/10/ A comparative analysis of selection schemes used in genetic algorithms David E. Goldberg Kalyanmoy Deb.
Genetic Algorithms An Example Genetic Algorithm Procedure GA{ t = 0; Initialize P(t); Evaluate P(t); While (Not Done) { Parents(t) = Select_Parents(P(t));
Introduction to AI (part two) Tim Watson G6.71
Estimation of Distribution Algorithms Ata Kaban School of Computer Science The University of Birmingham.
Theory Chapter 11. A.E. Eiben and J.E. Smith, Introduction to Evolutionary Computing Theory Overview (reduced w.r.t. book) Motivations and problems Holland’s.
COMP305. Part II. Genetic Algorithms. Genetic Algorithms.
Hierarchical Allelic Pairwise Independent Function by DAVID ICLĂNZAN Present by Tsung-Yu Ho At Teilab,
Genetic Algorithm for Variable Selection
COMP305. Part II. Genetic Algorithms. Genetic Algorithms.
Chapter 14 Genetic Algorithms.
Theory Chapter 11. A.E. Eiben and J.E. Smith, Introduction to Evolutionary Computing Theory Overview Motivations and problems Holland’s Schema Theorem.
Genetic Algorithms Sushil J. Louis Evolutionary Computing Systems LAB Dept. of Computer Science University of Nevada, Reno
Tutorial 1 Temi avanzati di Intelligenza Artificiale - Lecture 3 Prof. Vincenzo Cutello Department of Mathematics and Computer Science University of Catania.
Mathematical Models of GAs Notes from * Chapter 4 of Mitchell’s An Intro. to GAs * Neal’s Research CS 536 – Spring 2006.
Differential Evolution Hossein Talebi Hassan Nikoo 1.
Genetic algorithms. Genetic Algorithms in a slide  Premise Evolution worked once (it produced us!), it might work again  Basics Pool of solutions Mate.
© Negnevitsky, Pearson Education, Lecture 11 Evolutionary Computation: Genetic algorithms Why genetic algorithm work? Why genetic algorithm work?
Theory A.E. Eiben and J.E. Smith, Introduction to Evolutionary Computing Chapter 11 1.
Computing & Information Sciences Kansas State University Friday, 21 Nov 2008CIS 530 / 730: Artificial Intelligence Lecture 35 of 42 Friday, 21 November.
林偉楷 Taiwan Evolutionary Intelligence Laboratory.
A Brief Introduction to GA Theory. Principles of adaptation in complex systems John Holland proposed a general principle for adaptation in complex systems:
Schemata Theory Chapter 11. A.E. Eiben and J.E. Smith, Introduction to Evolutionary Computing Theory Why Bother with Theory? Might provide performance.
Estimation of Distribution Algorithms (EDA)
10/12/20151 V. Evolutionary Computing A. Genetic Algorithms.
CS Machine Learning Genetic Algorithms (II).
CS 484 – Artificial Intelligence1 Announcements Lab 3 due Tuesday, November 6 Homework 6 due Tuesday, November 6 Lab 4 due Thursday, November 8 Current.
Theory Chapter 11. A.E. Eiben and J.E. Smith, EC Theory, modified by Ch. Eick Overview Motivations and problems Holland’s Schema Theorem – Derivation,
An Introduction to Genetic Algorithms Lecture 2 November, 2010 Ivan Garibay
1 Machine Learning: Lecture 12 Genetic Algorithms (Based on Chapter 9 of Mitchell, T., Machine Learning, 1997)
Genetic Algorithms Siddhartha K. Shakya School of Computing. The Robert Gordon University Aberdeen, UK
1 Chapter 14 Genetic Algorithms. 2 Chapter 14 Contents (1) l Representation l The Algorithm l Fitness l Crossover l Mutation l Termination Criteria l.
Kansas State University Department of Computing and Information Sciences CIS 732: Machine Learning and Pattern Recognition Friday, 16 February 2007 William.
GENETIC ALGORITHMS.  Genetic algorithms are a form of local search that use methods based on evolution to make small changes to a popula- tion of chromosomes.
Genetic Algorithms José Galaviz Casas Facultad de Ciencias UNAM.
Why do GAs work? Symbol alphabet : {0, 1, * } * is a wild card symbol that matches both 0 and 1 A schema is a string with fixed and variable symbols 01*1*
Edge Assembly Crossover
MAE 552 Heuristic Optimization Instructor: John Eddy Lecture #12 2/20/02 Evolutionary Algorithms.
Genetic Algorithms Abhishek Sharma Piyush Gupta Department of Instrumentation & Control.
Chapter 12 FUSION OF FUZZY SYSTEM AND GENETIC ALGORITHMS Chi-Yuan Yeh.
2/29/20121 Optimizing LCLS2 taper profile with genetic algorithms: preliminary results X. Huang, J. Wu, T. Raubenhaimer, Y. Jiao, S. Spampinati, A. Mandlekar,
5. Implementing a GA 4 학습목표 GA 를 사용해 실제 문제를 해결할 때 고려해야 하는 사항에 대해 이해한다 Huge number of choices with little theoretical guidance Implementation issues + sophisticated.
Recent Research about LTGA Lecturer:Yu-Fan Tung, R Advisor:Tian-Li Yu Date:May 8, 2014.
Diversity Loss in General Estimation of Distribution Algorithms J. L. Shapiro PPSN (Parallel Problem Solving From Nature) ’06 BISCuit 2 nd EDA Seminar.
Sporadic model building for efficiency enhancement of the hierarchical BOA Genetic Programming and Evolvable Machines (2008) 9: Martin Pelikan, Kumara.
GENETIC ALGORITHMS Tanmay, Abhijit, Ameya, Saurabh.
1 Chapter 3 GAs: Why Do They Work?. 2 Schema Theorem SGA’s features: binary encoding proportional selection one-point crossover strong mutation Schema.
An Introduction to Genetic Algorithms Lecture 2 November, 2010 Ivan Garibay
Why do GAs work? Symbol alphabet : {0, 1, * } * is a wild card symbol that matches both 0 and 1 A schema is a string with fixed and variable symbols 01*1*
Something about Building Block Hypothesis Ying-Shiuan You Taiwan Evolutionary Intelligence LAB 2009/10/31.
Artificial Intelligence By Mr. Ejaz CIIT Sahiwal Evolutionary Computation.
By Ping-Chu Hung Advisor: Ying-Ping Chen.  Introduction: background and objectives  Review of ECGA  ECGA for integer variables ◦ Experiments and performances.
Genetic Algorithms And other approaches for similar applications Optimization Techniques.
Advanced AI – Session 7 Genetic Algorithm By: H.Nematzadeh.
مروری بر مفاهيم و کابردها
Chapter 14 Genetic Algorithms.
A comparative analysis of selection schemes used in genetic algorithms
Dr. Kenneth Stanley September 13, 2006
C.-S. Shieh, EC, KUAS, Taiwan
Chandimal de Silva and Joe Suzuki Osaka University, Japan
Genetic Algorithms Chapter 3.
SCHEMATA THEOREM (Holland)
New Crossover Scheme for Parallel Distributed Genetic Algorithms
Theory Chapter 11.
Presentation transcript:

Presenter : Tsung-Yu Ho

Review Mixing Problem Categories Crossover as a Mixer (√) Crossover as a Innovator (√) Crossover as a Disrupter All in one Crossover as a Disrupter Conclusion

Goldberg, Thierens, and Deb, 1993 “Toward a Better Understanding of Mixing in Genetic Algorithms.” Control Map How well “fixed crossover operators” solve ? GA-easy problem GA-hard Problem Before design a better operator Studies on existed crossover operators.

Four Crossover Categories (Different Model Problem) Crossover as a Mixer Crossover as an Innovator (BB) Crossover as a Disrupter (BB) All in one model (Innovator and Disrupter) What kind of crossover operators can solve problem well? Fixed operators. Other operators.

Robbin’s equilibrium (1918) & Geiringer’s Theorem (1944) Equilibrium distribution (Biology) Predicted equilibrium in GA Linkage disequilibrium (Christiansen, 1989) Rate of convergence to equilibrium (Rabani, 1998) Relaxation Time Uniform crossover : O( ln l ) One-Point crossover : O( l ln l ) Pr Ű gel-Bennett (2001) prove the result.

GA-Easy Problem Goldberg, Thierens, and Deb, 1993 Mixing time Control Map S and Pc Fixed crossover operators are good. GA-Hard Problem BB mixing. Two BB mixing m BB nixing Sweet-spot shrinks exponentially. Population size grows exponential.

Focus on Schemata Disruption Motivation from schema theorem. Studies suggest schema theory should be obeyed. Satisfy schema theorem does not guarantee BB mixing. Schemata Disruption models are useful to compare different crossover operators.

Syswerda (1989) Schema survive for analysis of fixed crossover operators. Spears and De Jong (1991) Multi-point crossover. De Jong and Spears (1992) The effect of crossover operators on population size. Goldberg and Sastry (2001) Satisfy Schema theory issue.

Syswerda, “Uniform crossover in genetic algorithms.”, 1989 Analyzed schema survival rates. Use empirical result to suggest that uniform crossover outperformed one-point crossover and two-point crossover in the most case.

Spears and De Jong, “An Analysis of Multi-Point Crossover.” 1991 Analyzed multi-point crossover to compare with uniform crossover. Uniform has higher schema disruption rate. They suggest the disruption has a positive role in balancing the exploration and exploitation.

Sampling disrupter is important for population diversity when population become homogeneous. Crossover Productivity The property of crossover for diversity. Crossover Productivity is easy to measure.

De Jong and Spears(1992) The effect of crossover operators on population size Small population sizes Uniform performs better. Large population sizes Two-point performs better. But, they did not give any analytical framework.

Goldberg and Sastry, “A Practical Schema Theorem for Genetic Algorithms Design and Tuning.”, Control Map Selection Pressure. Crossover Probability. Show obeying schema theorem can not guarantee BB mixing.

Schema Theorem (Holland, 1978; De Jong, 1975) Proportionate selection and one-point crossover. Simplified Practical Schema Theorem Goldberg & Deb, 1991

Large selection pressure (Sp -> ∞) BB growth ensure even if crossover is fully disruptive Small crossover probability ( Pc -> 0) BB growth is ensured for any selection pressure Easiest way to obey schema theorem Does not guarantee mixing.

8-bit deceptive trap function Single building block Global optimum : Local optimum :

Mixer Models Uniform Crossover is suggested. Disrupter Models n-point crossover is suggested. Innovation Models Fixed Crossover operators are not enough. All in one Models Does not help in GA design.