A VISUAL GENETIC ALGORITHM TOOL

Abstract

This dissertation addresses on developing a Visual Genetic Algorithm Tool for an optimization problem. Genetic algorithms are a part of evolutionary algorithms and are based on nature’s evolutionary process. Solutions are able to reproduce among themselves to create fitter individuals contributing towards the optimal solution. The attractiveness of Genetic Algorithm lies in the fact that it is capable of searching good solutions from a large search space. The optimization problem that this dissertation will visually address is the Traveling Salesman Problem, defined as a NP-Complete class of problems. Many algorithms have been designed to solve the problem and now this dissertation considers genetic algorithm as a method to solve the Traveling Salesman Problem visually. The class of algorithms used traditionally to solve the Traveling Salesman Problem is defined as approximation algorithms. The visual tool is able to clearly show the evolution of the chromosomes and the effects of genetic algorithm towards obtaining the optimal solution. In addition, the visual tool is able to graphically visualize the improvements made to the tour route optimization and a graph that documents the tour distance throughout the optimization process. The tool is then used solve the Traveling Salesman Problem and the results documented ensuring that the tool was proper designed and implemented.