A computational view of group problem saving

A problem arising out of group decision support research is to develop a method to specify and represent the important phenomena comprising the events of the participants (humans and machines). The authors propose a solution in the form of a computational model of group problem solving derived from a view of individual problem solving. They view group problem solvers from the perspective of their capacities to function as intelligent agents; that is, they may be either humans or machines. They argue that the group tasks in which such intelligent agents engage, such as `negotiation´ or `decision making´, are actually specific forms of problem solving, and that the cognitive mechanisms underlying `group problem solving´ are precisely the same as those comprising individual problem solving. One mechanism central to the model is that of learning; therefore, the authors discuss the role learning plays in group communication and coordination. In order to test one aspect of the group model, they conducted an experiment in knowledge sharing among an assemblage of artificial-intelligence scheduling agents. The results indicate that knowledge sharing can be quite beneficial, but that the effects vary according to specific task environment experiences