Informational Openness Enhances Collaborative Decision-Making in Groups of Autonomous Agents: A Cognitive Agent-Based Study

Collaborative decision making is central to the organization of society. Juries decide cases; boards and executive teams make strategic decisions; and voters elect government officials. It is common to think of such groups as decision making entities. We say that juries deliberate; that boards rule; and that electorates decide. But this language is imprecise. Real decision processes do not occur within any group as an abstract entity. Collaborative decision making always happens within and between individual group members. There is a rich body of research focused on how individuals decide together in groups. Most of this research assumes that individuals are already committed to the group they are in. This leaves open the question of how individuals decide to join decision making groups in the first place. We present an agent-based model of collaborative decision making designed to address this question. We develop this model in the context of existing research on task learning and intra-group communication. We demonstrate that collaborative decision making is done best when it is done by groups that are informationally, as well as structurally, open.