Explanatory Analysis of Parallel Genetic Algorithm for Traveling Sales Man Submitted by.pdfVIP

Explanatory Analysis of Parallel Genetic Algorithm for Traveling Sales Man Submitted by Juha Sukki Fahad Rizvi Table of Content 1. Introduction 3 2. Classification of GA 3 3. Multiple Population Parallel GA 4 4. Steps in Genetic Algorithm 4 5. Dividing the Population in Parallel GA 4 6. Network Topology 5 7. Migration 5 8. Conclusion 6 9. Reference 6 1. Introduction In our presentation we discussed about Parallel Genetic Algorithms, their application on Traveling Sales Man (TSP) problem. We also discussed how to general genetic algorithms could be made parallel. We were given a challenge to explain one of such parallel genetic algorithm for TSP. 2. Classification of GA Parallel Genetic Algorithms can be classified into three types. • Global Single Population Master Slave: In them there is one single large population and just like non-parallel GA. In it the evaluation of the fitness function is divided on several computers. A brief introduction was presented in the presentation. • Single Population Fine Grained: These are suited for massively parallel architectures. It consists of one spatially structured population. The mating/ cross over is done only in a limited area called neighborhood. These neighborhoods can be overlapping. The basic idea is to have one element/chromosome for each processor. A brief introduction was also presented in the presentation. • Multiple Population Coarse Grained: These are genetic algorithms that are different from the first two