A framework for coordinating decisions in a system with several decision makers and uncertainty

Recent work has highlighted the potential losses in performance that are likely to appear in a decentralized control system when decision makers are uncertain about (or disagree about) which model of the system should be used. At best, this may simply reduce any incentive to coordinate activities, but it is more likely to cause the system´s performance to deteriorate. In practice, performance does seem to be quite sensitive to such model errors. A framework is presented in which decision makers jointly choose their model and their strategies so that the gains from coordination are not jeopardized by model misspecifications and/or disagreements. This framework leads to an alternative bargaining process and more successful coordination