Title: DECISION FROM INDECISION IN MULTI-AGENT MULTI-OPTION DYNAMICS
How does a group of agents break indecision about a set of options? Many biological and artificial multi-agent systems, from honeybees and bird flocks to bacteria and humans, face indecision when choosing between options in situations in which the fitness or even the survival of the group is at stake. Here, we develop a mathematical theory to study decision from indecision. Our approach is grounded in network bifurcation theory. We model decision from indecision as synchrony-breaking in influence networks in which each node is the value assigned by an agent to an option. We show that there are three generic types of value patterns emerging at synchrony-breaking bifurcations for influence networks: deadlock, consensus, dissensus. Deadlock and consensus value patterns are predicted by the symmetry of the influence networks. Conversely, we show that there are many “exotic” dissensus value patterns; that is, patterns that are not predicted by network symmetries, but are predicted by network architecture. Numerical simulations of a novel influence network model illustrate our theoretical results.