In this research project, an innovative methodology is developed for analysis, design and simulation of high-order complex nonlinear dynamic systems having an underlying multiple time- scale structure. A general approach is formulated for time-scale identification, efficient numerical simulation, and reduced- order model development for high-order nonlinear systems encountered in the context of designing machines, processes, structures, vehicles and robotic manipulators. The methodology is based upon the use of Computational Singular Perturbation (CSP) technique, which is an iterative decoupling procedure originally developed for singularly perturbed linear systems. The application of this methodology has been successfully demonstrated for stiff problems in chemical kinetics. In this research, both initial and boundary value problems are addressed. Formulating system design problems in the framework of optimal control leads to Hamiltonian boundary-value problems which often have multiple time-scale structures. The modified and extended CSP method is used as a user friendly software-oriented general tool for such problems, and provides a significantly new capability for complex nonlinear system analysis and design.*** //
