Abstract Proposal: 97-33929 Northwestern University PI: Athanasius Ziliaskopoulos "CAREER: On-line Optimization and Control of Real-Time Transportation Systems" This CAREER award supports research concerned with the problem of on-line control and optimization of the operations of a real-time dynamic and stochastic transportation system. The motivation for this research is two-fold: to use to use new technological advances to improve the control and management of transportation operations; and to fill the gap created by the lack of algorithms that can use these technologies to improve the efficiency of transportation operations. In addition to the research problems initiated by the technological advances, several fundamental educational issues are raised: In general, students and practicing transportation engineers do not have sufficient technical background to fully appreciate the complex behavior of stochastic and dynamic systems and the knowledge to exploit the potential of real-time data. This is partly due to training that has traditionally been limited to the concepts of equilibrium, steady state average system behavior and determinism. Educators are faced with the challenge of developing innovative pedagogical material and methodologies that convey the fundamental principles of dynamic and stochastic systems, given the limited background of typical engineering students. The main research and educational objectives are: to analyze the operations of real-time transportation systems and identify their properties; to develop closed form analytical formulations by exploiting their special properties and to explore various exact and heuristic solution approached, optimality conditions for exact algorithms and performance bounds for heuristic approaches; and to develop new pedagogical tools that will enable students to understand the fundamental principles of dynamic and stochastic systems. To achieve these objectives, concepts from four major areas will be utilized: stochastic programming, stochastic optimal control, process control, and Markovian decision processes.
