| This work represents the design, analysis and solution, from the systems control perspective, of a complex, dynamic, and multi-constrained planning system for generating training plans for crew members of the NASA International Space Station. Various intelligent planning systems have been developed within the framework of Artificial Intelligence (AI), but these planning systems lack a rigorous mathematical formalism to allow a reliable and flexible methodology for design, modeling, and analysis of their performance in a dynamical, time-critical, and multi-constrained environment. Formulating the planning problem in the domain of discrete-event systems under a unified framework such that it can be modeled, designed, and analyzed as a control system will provide a self-contained theory for planning systems. This will also provide a means to certify various planning systems for operations in the dynamical and complex environments in space. This dissertation completes the development of such an intricate, large-scale, and representative mathematical formulation, as well as design and analysis, for intelligent control of a planning system for Space Station crew training. The planning system was developed as a control system, with the plant and the supervisory controller modeled as a Discrete Event System (DES) using Petri Nets, and with a planner that implemented an heuristic search algorithm (the A... |
