Memetic Algorithms By Anup Kulkarni08305045 Prashanth K08305006 Instructor

Memetic Algorithms By Anup Kulkarni(08305045) Prashanth K(08305006) Instructor: Prof. Pushpak Bhattacharyya

Overview Philosophy Behind Memetics Genetic Algorithm – Intuition and Structure Genetic Algorithm Operators Memetic Algorithms TSP Using Memetic Algorithm Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 2

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. Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 3

Biological Evolution • Natural Selection • Survival of The Fittest • Origin of New Species Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 4

Examples of Biological Evolution and Natural Adaptation Gills in Pisces 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 Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 6

Biological Evolution Meme. . !!! Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 7

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 Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 8

Examples of Meme Fashion Science Scientists sharing their thoughts Literature Latest trends are ideas of fashion designers 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 Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 10

Genes and Memes, where they differ Genes are pre-decided Genes are static through generations, memes can be changed! Memes allow improvement After learning language, we contribute to it through literature New heuristics to 8 -puzzle problem solved in class We use this property to improve genetic algorithms Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 11

Genetic Algorithm solves (typically optimization) problems by combining features of complete solutions to create new populations of solutions. applicable when it is hard or unreasonable to try to completely identify a subproblem hierarchical structure or to approach the problem via an exact approach. Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 12

Genetic Algorithm Initialize population Pop Evaluate Pop While not stop criterion do Select Parents from Pop Recombine Parents Evaluate Pop Return the best solution in Pop

Crossover 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 0000000 0001111 1110000 Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 14

Mutation ■ Purpose: to introduce new characteristics in the population by random modifications. before ■ after Explores the “neighborhood” of a solution. 1 1 1 1 1 0 1 1 1 mutated gene value

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

Solving the Traveling salesman problem with a Memetic Algorithm Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 17

Memetic Algo for TSP-representation Array pop stores population Size of pop=P No of cities=N Tour represented as 1234. . N Fitness function-cost of the tour Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 18

TSP - Crossover Distance Preserving Crossover d(p 1, p 2) = d(p 1, child) = d(p 2, child) d(x, y) = #edges not common in x and y Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 19

Distance Preserving Crossover Source: B. Freisleben et al, “New Genetic Local Search Operators for the Traveling Salesman Problem” Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 20

2 -OPT Search Delete any two edges Insert other two edges which will result in new 1 1 tour 6 2 3 2 6 3 5 5 4 4 Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 21

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

Performance Source: Slides of A. E. Eiben and J. E. Smith, Introduction to Evolutionary Computing Hybridisation with other techniques: Memetic Algorithms Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 23

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. Optimal solution=Genetic Algo + Local Search Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 24

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, 1 st edition, Addison-Wesley Longman Publishing Co. , 1989 pp 170 -174 B. Freisleben and P. Merz, New Genetic Local Search Operators for the Traveling Salesman Problem. In H. -M. Voigt, W. Ebeling, I. Rechenberg, and H. -P. Schwefel, editors, Proceedings of the 4 th Conference on Parallel Problem Solving from Nature - PPSN IV, pages 890 --900. Springer, 1996 S. Lin and B. W. Kemighan, An effective heuristic algorithm for the Traveling Salesman problem, Operation Research 21 (1973) 498516 Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 25

? Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 26

Thank you! Anup Kulkarni and Prashanth K, Dept of CSE, IIT Bombay 27