Models of consensus are used to manage multiple agent systems in order to choose between different recommendations provided by the system. It is assumed that there is a central agent that solicits recommendations or plans from other agents. That agent the n determines the consensus of the other agents, and chooses the resultant consensus recommendation or plan. Voting schemes such as this have been used in a variety of domains, including air traffic control. This paper uses an analytic model to study the use of consensus in multiple agent systems. The binomial model is used to study the probability that the consensus judgment is correct or incorrect. That basic model is extended to account for both different levels of agent competence and unequal prior odds. The analysis of that model is critical in the investigation of multiple agent systems, since the model leads us to conclude that in some cases consensus judgment is not appropriate. In addition, the results allow us to determine how many agents should be used to develop consensus decisions, which agents should be used to develop consensus decisions and under which conditions the consensus model should be used.
Download