Performance evaluation of ZEUS, Jade, and SkeletonAgent frameworks

Due the growing interest in the field of intelligent agents and multi-agent systems researchers have developed different toolkits. The aim of those toolkits, or frameworks, is to help the designers and engineers to build complex systems based on the agent concept. This paper presents a brief description of some of those frameworks: ZEUS, Jade and SkeletonAgent. These frameworks use their own agent architecture and other facilities like visual programming toolkits, documentation or reusable software libraries to facilitate the definition and development of multi-agent systems. The main aim of this paper is to compare the different frameworks in a common domain: to search news in several electronic newspapers. Because every one of those multi-agent toolkits use different features to build the whole multi-agent system, the behavior of any of those systems is expected to be different. The empirical evaluation measures the request time and the number of retrieved documents for the different systems. Finally, the paper discusses the conclusions for the previous experiments.