Introduction To Genetic Algorithm

2.1 INTRODUCTION

Genetic Algorithm was first developed by John Holland at the University of Michigan in the year 1980. Generally Genetic Algorithm’s are said to be random process, which provides an optimal solution from a given set of possible solutions by a executing a series of problems. For solving the optimization problems Genetic Algorithms performs intelligent technique at random search. Even though Genetic Algorithm process is said to be at random search, the results produced are not random.

Even though many variants are available for Genetic Algorithms, the working principle of Genetic Algorithm is not different for any type of problem depicting its robustness for problems in several domains. Based on the best fitness obtained from

These techniques may produce unpredictable results when there are multiple peaks and points in the search space. But, GA starts with multiple numbers of solutions and several points are explored in the search space in parallel. Thus the chance of getting stuck in the local and global optima is reduced. Search procedure is not a random process and the number of genetic operators applied to help this optimization process.

2.3 Basic Structure of Genetic Algorithm

In an effective Genetic Algorithm model, the first step is to design the solution as a bit vector or number of chromosome. The collection of chromosomes forms a population representing the search space in a available vector. Based on the problem the objective function is designed to find the fitness evaluation phase. Next generation begins after the evaluation of present individuals is done. This is done with the help of selection mechanisms and the genetic operators like Crossover and Mutation operators.

This procedure is repeated till the termination condition is reached or best optimal solution is identified.

The pseudo-code for the Standard Genetic Algorithm is as followed by the steeps:

1.Generate the initial population of ‘n’ Chromosomes by random

These parameters are used for evaluation. In biological terminology, each individual is called as chromosome and each parameter in the chromosome is called as gene. The process of representation of parameters in the form of structure is called as encoding. Generally, the search process of Genetic Algorithms is done on a collection of individuals or chromosomes called as population.

Commonly used chromosome representations in Genetic Algorithm are

1.Binary Encoding

2.Real Value Encoding

3.Permutation Encoding

4.Tree Encoding

2.4.1.1 Binary Encoding

In Binary Encoding the chromosomes are represented in the form zeros and ones, the binary digits. Many of the Genetic Algorithm implementations follow binary encoding because it is easy to implement and it can manipulate easily. Generally the bit strings are represented based on the availability of the individual, if the offspring is present in the analysis then it is represented as 1 otherwise it is