Towards an engineering of multi-agent eco-systems

Different approaches for engineering multi-agent systems have been proposed over time. However, there is a lack of attention towards the social aspects of multi-agent systems. In this paper we present an approach for engineering multi-agent systems, which considers multi-agent systems as digital ecologies. Different agent ecologies can work together and act as digital ecosystem. In such ecosystem the social aspects of the agents are well defined and all the subcomponents of the multi-agent system work together to achieve a fine integration of the whole ecosystem. Instead of focusing on internal model of software agents our methodology attempts to capture the dynamic and social behaviors of agents by focusing on the cooperation, coordination, negotiation and management attributes of the system. From implementation point of view, the proposed engineering methodology is grounded in our multi-agent framework and the concepts developed using this methodology can be mapped to implementation easily. The framework provides support for integration of web-services and uses peer-to-peer approach for resource discovering.