Engineering mobile agents

Abstract

As the mobile agent paradigm becomes of interest to many researchers and industries, it is essential to introduce an engineering approach for designing such systems. Recent studies on agent-oriented modeling languages have recognized the need for modeling mobility aspects such as why a mobile agent moves, where the agent moves to, when it moves and how it reaches its target. These studies extend existing languages to support the modeling of agent mobility. However, these fall short in addressing some modeling needs. They lack in their expressiveness: some of them ignore the notion of location (i.e., the "where") while others do not handle all types of mobility (the "how"). Additionally, they lack in their accessibility, as the handling of the mobility aspects is separated into multiple views and occasionally the mobility aspect is tightly coupled with the functional behavior specification. View multiplicity reduces the comprehensibility and the ease of specification, whereas the coupling with the functional behavior specification reduces the flexibility of deploying a multi-agent systems in different configurations (i.e., without mobility). In this paper, we address these problems by enhancing an expressive and accessible modeling language with capabilities for specifying mobile agents. We provide the details of the extension, then illustrate the use of the extended modeling language, and demonstrate the way in which it overcomes existing problems.