sBAM in solving Traveling Salesman Problem

The Traveling Salesman Problem (TSP) has already been solved in the semi-optimal manners using the numbers of different methods. Among them, genetic algorithms (GA) are pre-dominating. This paper presents a new approach to solve this problem using the Simplified Bi-directional Associative Memory (sBAM), a type of Artificial Neural Network. To get a comparative idea of its performance, the same problem has been solved using a genetic algorithm. In this paper, performance has been analyzed of a TSP by Genetic Algorithm (GA) and sBAM. Finally we proved that sBAM provide real time highly faster nearly optimal solutions than the genetic algorithm.

Boshir Ahmed is an Assistant Professor in the Department of Computer Science and Engineering at Rajshahi University of Engineering and Technology (RUET), Bangladesh. His research interests include Microprocessor Based System Design, Digital Communication Systems, Computer Networking, Neural Network, Digital System Design and Algorithms.