Testing Agent-Oriented Software by Measuring Agent’s Property Attributes

Abstract

Agent technology has been the subject of extensive discussion and investigation within the research community for several years, but it is perhaps only recently, it has seen any significant degree of exploitation in commercial applications. We could view Agent Oriented Software Engineering (AOSE) as a new programming paradigm that has evolved itself from Object Oriented Software Engineering (OOSE). AOSE has placed greater emphasis on agent characteristics such as the autonomy, learning, interaction, adaptability, reactivity, pro-activity etc. However, no paradigm will be completed if there were no form of measurement (metrics) to determine the efficiency and quality of its application. AOSE needs proper product metrics to enhance its stand as a lasting methodology in software engineering, just like its parent. For this reason, it is important to develop comprehensive measures of excellence to evaluate agent based software. No set of measures defining the overall quality of an agent has been developed to date. Our research is an elucidation to the need for a more encompassing definition of agent quality through metrics. We present a set of metrics that can be easily applied to a design that measures certain software quality characteristics of an agent oriented system. These metrics are “design and product” metrics for Agent Oriented system. They provide information regarding the ability of their design to match software quality. This paper first identifies the quality requirements for an agent system and secondly proposes a tool that has been developed to measure the agent metrics that quantifies and assists quality engineers to decide the efficiency and the degree of quality of Agent-oriented system.