Emergence of Multiagent Coalition by Leveraging Complex Network Dynamics

Emergence of a single coalition among self-interested agents operating on large scale-free networks is a challenging task. Existing approaches, both centralized and decentralized, suffer from high overhead costs to maintain network wide communiation. Furthermore, these works start by assuming a pre-established static complex network platform and then employ agents on the nodes of the network for mutual interactions. In other words, agents do not have the capability to form the network. These approaches use a given structure of the complex network with fixed properties and do not explore the other possible network configurations by varying the topological features. In addition to this, they do not use the network dynamics to control the dynamics of agent interactions. In this paper, we present a decentralized game-theoretic approach to this single coalition emergence problem that limits agent communications only to their immediate neighbors. We enable the agents to choose their interaction partners to form a dynamically growing SF network and show that this network formation process facilitates that emergence phenomenon. We perform an extensive computational study by varying some topological properties over two models of the scale-free network. The novel insight that we gain is that by increasing degree-heterogeneity and clustering-coefficient, implemented through agents’ partner selection strategy, we could enhance the likelihood of the emergence of a sustained single coalition over various types of scale-free networks.