Research Objectives and Approaches: The objective of this project is to advance the understanding, principles, and applications of a new learning framework that integrates optimization and prediction based on adaptive dynamic programming. The approach is to integrate a reference network into the adaptive dynamic programming design to link optimization and prediction by providing the internal reinforcement/goal representation. The proposed architecture can be designed in a hierarchical way, in which different levels of internal reinforcement signals can be developed to represent multi-level goals for an intelligent system.
Intellectual Merit: This project proposes a new reference network to the adaptive dynamic programming design to integrate optimization and prediction capabilities for brain-like intelligence research. This structure provides a new approach to adaptively build the internal reinforcement signal rather than crafted by hand. The underlying principle and mathematical foundation of using internal reinforcement/goal representation could potentially bring cognitive optimization and prediction closer to reality.
Broader Impacts: This research has the potential to develop innovative techniques for a wide range of applications, such as power grid control and communication/network systems. In addition to the benefits for academia and industry, this project will also benefit national and government interests as well, such as providing technical innovations for the U.S. energy security and sustainability objective (e.g., smart grid). The integrated research and education plan will not only promote excellence in education at different levels, but it will also enrich the participation of students with different backgrounds and encourage women and minority students to enter this field.
