This proposal describes the PIs early career development plan for integrated research and education centered around the unified theme of cooperative systems design by a stochastic games approach.
Intellectual Merit: A cooperative system consists of a number of interacting (semi-)autonomous agents where each individual agent makes local decisions based on the local information available to optimize its own local objective. The purpose of designing such a cooperative system is to optimize a global objective which may be thought of as system designers objective. In other words, system designer seeks the optimization of its own objective through multiple agents that seek the optimization of their own individual self-objectives. Examples of cooperative systems include multi-vehicle search and rescue operations, multi-vehicle search and target assignment for military mission planning, multi-sensor deployment for anti-submarine warfare, and cooperative multi-user MIMO signaling in wireless communication systems. Stochastic games, which generalize Markov decision processes to multiple decision makers, emerge as the most natural framework to study cooperative systems. The main goals of the proposed work can be summarized as follows: Developing theoretical and computational tools for the analysis and synthesis of cooperative systems within the framework of stochastic games Demonstrating the value of the stochastic games approach in cooperative systems design to industry through new applications. Introducing the game theoretical methods for engineering to graduate and undergraduate students through new course development and application based multidisciplinary research projects. Making the subject accessible to a broad audience including traditionally under-represented groups and engineering community affiliated with the local industry throughout Hawaii islands using distance and web-based learning tools some of which are already available within University of Hawaii system. The basic premise of the proposed work is that the PIs postdoctoral research at UCLA, which led to significant insights to design of cooperative systems within the context of finite games, has the potential of breeding more breakthrough ideas to push the state of the art knowledge in cooperative systems to new frontiers by accomplishing the above goals.
Broad Impact: Cooperative systems design is truly an inter-disciplinary topic that connects to a many real-world applications in such diverse areas as national security, telecommunications, transportation networks, and power systems. As such, the research and education activities delineated in this proposal would potentially make a positive impact on the lives of many citizens. More specifically, the university-industry collaborations are expected to help improve nations economy and create employment while contributing to national security. Another specific benefit of such collaborations in the area of transportation systems would be congestion reduction in heavily congested traffic systems using the design methods developed for cooperative systems as a result of the proposed research activities. The proposed educational activities are expected to make the subject of this proposal accessible to a broad audience including traditionally under-represented groups and engineering community affiliated with the local industry throughout Hawaii islands using distance and web-based learning tools. This would create more educational opportunities to those groups that have limited access to education. An interactive website that will be developed as part of this project is expected to increase public knowledge about game theory and its applications to cooperative systems design. Moreover, such a web-site is expected to intrigue curious K-12 students and help them in their career decisions by illustrating the significance of cooperative systems design through simulations and animations. A1
