Fundamentals Of Genetic Algorithm, Simulated Annealing And Tabu-Search

Introduction

Genetic algorithm or GA is a stochastic search technique which is based upon the mechanism of natural selection and genetics. This can be regarded since an intelligent search methodology which needs a domain-specific knowledge to handle a combinatorial optimization problem. The key aspect of the genetic algorithm is that this starts its search from a population of points here the objective function is calculated at all point in the search space as one at a time. Genetic algorithm mimics the process of natural evolution by combining the survival of the fittest in between solution structures along with a structured, until randomized, information exchange and creates offspring as Goldberg in 1989. The offspring displaces weak solution during all generation.

A candidate solution is presented either by a decimal or binary sequence of number and is termed as chromosome or string. A chromosomes potential as a solution is find out by its fitness function, which need the evaluation of chromosomes along with respect to the objective function of the optimization problem beneath consideration. A wise selection of chromosomes constitutes a population and at a specified time, this population makes a generation. The population size one time fixed remains similar from generation to generation and has a major impact on the performance of GA. This operates on a generation concept and mainly performs three operations name is:

Reproduction

Selection of copies of chromosomes is proportional to their fitness value

Crossover

An exchange of sections of the chromosomes

Mutation

Randomly modification of the chromosomes

The outcome of above three operations is termed as offspring or children and forms the next generation population. This process is continued for a desired number of generations, generally upto a point where the system converges to an important well performing sequence. Programme 1 demonstrates the flowchart of basic genetic algorithm such can be applied in combinatorial optimization problems.