Goal-based Multi-agent Collaboration Community Formation: A Conceptual Model

Multi-agent systems mostly consist of selfinterested agents, which behave selfishly to achieve their goals. However, their limited capabilities and knowledge, environment’s constrained resources and shared goals lead them to a collaborative goal achievement process. Based on previous findings, assuming fixed organizational structures is not practical solution for dynamic multi-agent systems. Therefore, agents need to acquire an understanding of the other agents’ goals and policies, update their current perception of their environment constantly and form/reorganise an effective community for any instances of a collaboration process. This paper summarizes a Collaboration Community Formation Model (CCFM) introduced in [Golpayegani et al., 2016] in a conceptual level. This model describes the fundamental structural elements of a dynamic collaborative community. We also introduce an agent model which describes agents’ internal components that allows agents to model their dependencies using goal reasoning and self-organise their collaboration community and select their potential collaborators amongst the numerous agents in the system.