Information Induced Multimodel Solutions in Multiple Decisionmaker Problems

This paper is concerned with modeling and control strategy interaction in a multimodel context. The role of the observability structure in multiple decisionmaker (DM) problems is examined. A procedure is developed for generating multimodel solutions based on certain deterministic information patterns. This is achieved by first explicitly identifying the state space structure induced by the observation sets of the DM's, and then overlapping appropriately the input space structure of each DM. Such a representation is used to identify the class of admissible strategies which generate multimodel solutions. Conditions, which depend on the information pattern, are obtained under which these multimodel solutions admit partial noninteraction among the DM's.