The Design Of New Learning Automata For Problem Graph Coloring

Abstract

Graph coloring issue, is one of satisfied existing constraints issues in the literature of artificial intelligence. Coloration apical includes assigning color to node graph so that any two adjacent vertices are isochromatic. The minimum number (colors numbers) that we assign to these graphs for coloring are called number of color. This issue is from the group of very difficult issue, NP – complete. Given the importance of graph coloring issue and its many uses, many algorithms suggested finding allowed coloration in graph. Among these can be noted in, exact algorithms, distributed algorithms, parallel algorithms, approximation algorithms and heuristic algorithms, …The concept of learning Automata at first was introduced by Tstlyn. He was interested in modeling the behavior of biological systems, and definite automata worked in a random environment, introduced as a model for learning. The aim of this research is to present new algorithm on the basis of learning automata to color with accuracy and high speed and the ability to learn graph vertices. The proposed method also has transfer chart and individual performance and this method was examined on the graph with low vertices and high vertices and medium vertices, on the bottom, a number of works steps and total dyes used for coloring of specific graph with optimization algorithms were matched. Evaluation results show high accuracy, speed and its performance of the proposed method is superior to other optimization methods.