440 likes | 887 Views
Elena Marchiori. Memetic Algorithms. Bio-Inspired Algorithms Course. Topics: Ant Colony Optimization (ACO) Memetic Algorithms Multi-objective Optimization Structure: Theory: learn above mentioned topics Practice: work in groups on application-oriented problems
E N D
Elena Marchiori Memetic Algorithms
Bio-Inspired Algorithms Course • Topics: • Ant Colony Optimization (ACO) • Memetic Algorithms • Multi-objective Optimization • Structure: • Theory: learn above mentioned topics • Practice: work in groups on application-oriented problems • See http://www.cs.ru.nl/~elenam/BIA.html
Overview of today’s sample lesson on memetic algorithms Introduction Genetic Algorithm – Intuition and Structure Genetic Algorithm Operators Memetic Algorithms MAXclique Using Memetic Algorithm
Genes and biological evolution A gene is a unit of biological information transferred from one generation to another. Genes determine our physical traits, what you look like, what you inherit from either one of your parents.
Biological Evolution • Natural Selection • Survival of The Fittest • Origin of New Species
Examples of Biological Evolution and Natural Adaptation • Frog Skin • Hollow Bones in Birds • Biological Evolution of Human • Characteristic Thumb • Erect Vertebral Column • Lower Jaw
Biological Evolution Cultural Evolution..?? Source: www.wikipedia.org
Biological Evolution Meme..!!!
Meme • “the basic unit of cultural transmission, or imitation” - Richard Dawkins • “an element of culture that may be considered to be passed on by non-genetic means” - English Oxford Dictionary
Examples of Meme • Fashion • Latest trends are ideas of fashion designers • Science • Scientists sharing their thoughts • Literature • Novel, poetry • Music • Even birds are found to imitate songs of other birds!!!
Genes and Memes, where they are similar Genes propagate biologically from chromosome to chromosome Memes propagate from brain to brain via imitation Survival of fittest in meme Concept of God is survived though no scientific evidence is present
Genes and Memes, where they differ Genes are pre-decided Genes are (almost always) static through generations, memes can be changed! Memes allow improvement After learning language, we contribute to it through literature New heuristics to search problems We use this property to improve algorithms
Genetic Algorithm solves (typically optimization) problems by combining features of candidate solutions to create new populations of candidate solutions. applicable to complex optimization problems, e.g., when it is hard to tackle the problem via an exact search approach.
Initialize population Pop Evaluate Pop Select Parents from Pop Recombine Parents Genetic Algorithm While not stop criterion do Evaluate Pop Return the best solution in Pop
Crossover 0 0 0 1 1 1 1 1 1 1 0 0 0 0 • Purpose: to combine features of feasible solutions already visited in order to provide new potential candidate solutions with better objective function value. • Mechanism that restarts the search by “exploring” the space “between” solutions. parents offspring 0 0 0 0 0 0 0 1 1 1 1 1 1 1
Mutation 1 1 1 1 1 1 1 before 1 1 1 0 1 1 1 after mutated gene value • Purpose: to introduce new characteristics in the population by random modifications. • Explores the “neighborhood” of a solution.
Memetic Algorithm Initialize population Pop Optimize Pop(Local search) Evaluate Pop While not stop criterion do Select Parents from Pop Recombine Parents Optimize Pop(Local search) Evaluate Pop Return the best solution in Pop
GA main characteristics 1 2 • Representation: • 12345678 • 01011000 • Fitness: number of nodes of the sub-graph (=3 for the above sub-graph) • Selection: roulette-wheel (fitness proportional) • Elitism: copy best individual to next generation • Replacement rule: selects two best among parents and offsprings • Crossover: uniform • Mutation: swap values of two randomly selected genes 4 3 5 7 6 8
The Local Search Module • Candidate solution is any subgraph in the population • Perturb the sub-graph • Extract a clique • Expand the clique
Perturb added removed
Extract a clique added perturb extract removed removed
Enlarge to maximal clique perturb extract enlarge added added removed removed
Conclusion A genetic algorithm promises convergence but not optimality. But we are assured of exponential convergence, possibly at different optimal chromosomes. Do very well in identifying the regions where those optima lie. Efficient and effective search: Genetic Algorithms + Local Search
References R. Dawkins, “The Selfish Gene – new edition”, Oxford University Press, 1989 pp 189-201. David E. Goldberg, Genetic Algorithms in Search, Optimization and Machine Learning, 1st edition, Addison-Wesley Longman Publishing Co., 1989 pp 170-174. E. Marchiori . A Simple Heuristic Based Genetic Algorithm for the Maximum Clique Problem.ACM Symposium on Applied Computing ( SAC98), pp. 366-373, 1998.