Utilizing incomplete block designs in evaluating agent-oriented software engineering methodologies

In the past couple of decades, agent-oriented technology has been arisen in order to assist in developing intelligent software that is able to solve challenging problems. More than twenty agent-oriented software engineering methodologies (AOSE) are presented in literature. Nevertheless, the immaturity of this emerging technology made developers run into extreme difficulties when deciding which methodology can best fit a prospective application. A little work has been conferred to compare and evaluate small numbers of agent-oriented methodologies. However, such related work lacks the reliable framework that can be effectively implemented; most of the proposed approaches are not capable to provide sufficient knowledge to support the decision of selection. In this paper, we present a more reliable framework that adopts statistical techniques to evaluate AOSE methodologies and come up with a set of metrics that help in selecting the most appropriate methodology, or assembling more than one, to accommodate the anticipated characteristics of the desired application