Investigating Tabu Search for Web Effort Estimation

Abstract

Tabu Search is a meta-heuristic approach successfully used to address optimization problems in several contexts. This paper reports the results of an empirical study carried out to investigate the effectiveness of Tabu Search in estimating Web application development effort. The dataset employed in this investigation is part of the Tukutuku database. This database has been used in several studies to assess the effectiveness of various effort estimation techniques, such as Linear Regression and Case-Based Reasoning. Our results are encouraging given that Tabu Search outperformed all the other estimation techniques against which it has been compared.