Modelling resource allocation in open embedded systems

We consider the problem of deciding resource allocation policies in open embedded systems, where the system components have to form opportunistic alliances, share partial knowledge and collectively agree a policy that is congruent with the state of the environment in which the system is embedded. This problem arises in a number of systems and applications, including sensor networks, virtual organisations, cloud computing, swarm robotics, and infrastructure management. In this position paper, we outline a model for addressing this problem, and overview some of its features for which logical and epistemological issues concerning the dynamics of beliefs in a multi-agent system turn out to be relevant.