Data fusion using Tabu crossover genetic algorithm in information retrieval

Abstract

This article explores the odd and even point crossover based Tabu Genetic Algorithm. The search optimization tools equipped with exploration and exploitation operators. Those operators assist the optimization tools for finding the optimal solution. Few problems demand vigorous exploration and minimal exploitation. The vigorous exploration needs some specialized operators, which is capable of carrying out the task. In this article, we explore one such possible operator using odd and even point (OEP) crossover. The resultant hybrid GA namely OEP crossover based Tabu GA has two tuning factors namely tenure period and OEP crossover probability (Podd). The tenure period may be a single entity or a group of entities. However, Podd is single, as the tenure period is involved with group of entities, it demands some fine tuning. The fine tuning may alter the proportion of exploration and exploitation. Hence, we proposed a method for selecting the tenure period. The proposed exploration operator and the method for fixing the tenure period has been tested over the data fusion problem in information retrieval. The results look promising.