Indian Journal of Science and Technology

Abstract

In all the type of physical networks, optimization is one of the important factors in the minimization of the cost. This paper presents an approach to connect the distributed network locations in the form of single network tree using the concept of genetic algorithm in such a way that total covered distance should be minimum and the connectivity of the node in the given preference should also be maintained. Connectivity of the node can be defined as DCMST problem which comes under the category of NP problems. Inclusion of additional constraints makes it NP hard problem. Approximation algorithms have no optimal solutions so far that is why metaheuristic approach Genetic Algorithm is applied and satisfactory results have been found. Current research has considered the size or network from 10 nodes to 500 nodes of network with fully connected. Population range and generation range has been considered from 10 to 5000 with the genetic operators with the uniform crossover and mutation insertion. This work has shown that population variation with the medium range and minimum number of generation can produce the better results in the comparison of extremely large size population or maximum number of generations. Optimization of the network is a real life worldwide useful problem and it has many applications from the data communication, VLSI design to logistic transition. 
Keywords: Communication, Degree Constraint, Genetic Algorithm, Route Optimization, Spanning Tree