Abstract: Deploying a large number of mobile agents in scenarios where agents migrate frequently and/or exchange messages frequently requires methods for message delivery that are adequate to these specific situations. Deciding on which message delivery model to use, and whether a newly developed model is better than existing ones, may be difficult without an experimental testbed for comparison. This paper presents a framework for the comparison of message delivery models dedicated to mobile agent systems. The framework allows the generation of large, difficult scenarios, in which different methods may be evaluated side-by-side, revealing trade-offs between success rate, delivery time, and resource consumption. The architecture of the framework is designed to quickly integrate new models and to allow the direct deployment of a model implementation in real-life applications. As validation, we have integrated the implementation of several well-known delivery models and made comparisons between these models, from different points of view.
Loading