The organizational structure in collaborative information gathering

Collaborative information gathering (CIG) is an approach to solving the problem of Internet information overload by organizing cooperation among users. In the paper a multiagent perspective is taken to analyze CIG systems in an organizational level. We provide a multiagent model to describe the organizational structure of CIG systems and then validate it by comparison between the actual CIG systems and the model itself. In our model, the distribution entropy of resource is used to describe the task environment. The idleness rate and error rate are used to measure individual performance. The collective performance is scaled by the percentage of gathered resource. By utilizing this model in a series of simulations, three typical organizational structures are compared in terms of collective performance and performance stability. The results of the simulations lead to several conclusions about how to organize efficient collaborative information gathering.