9410313 Chong This proposal deals with issues in parametric and nonparametric decision and control problems for Discrete Event Systems (DESs). DESs are dynamical systems that are driven by the occurrence of physical events, and arise in virtually all man-made systems, including manufacturing systems, communication networks, distributed computer systems, logistic systems, and vehicular traffic. The proposal considers performance oriented DES models, and focuses on the problem of performance optimization. Methods for optimizing DESs have received significant interest in recent years, and show tremendous promise for the development of efficient ways to optimize large-scale complex man-made systems. An objective of the study is to couple perturbation analysis methods with optimization algorithms (both parametric and nonparametric), with the aim of gaining further rigorous understanding into when and why they work, and developing algorithms and tools that will be useful in DES decision and control problems. This work emphasizes applications in manufacturing systems, communication networks, and real-time computer systems. ***