Cluster Based Multi Agent System for Breadth First Search

Abstract

Tree searching problems plays vital role in artificial intelligence and the most of the tree search algorithms have been implemented using structured or object oriented programming. The sequential breadth first search algorithm starts traversing from a source or starting node to neighbour nodes which are adjacently attached to source node in the graphs. It is based on the layer wise exploring the neighbour nodes. The design of multi agent system for searching algorithms is one of the solutions for the algorithm efficiency in graph algorithms. In this paper, the design of multi agent system for breadth first search is presented with implementation detail. The proposed design divides the whole graph into several clusters by using bi-connected region method and assigns the agents to each clusters to perform the breadth first search in each cluster sequentially. The performance measure of the proposed method reveals the competitive results in terms of the sequential breadth first search algorithm.