Simulation Optimization
Stanisław Raczyński · Modeling and Simulation · 2006
The term Simulation Optimization is used for optimization problems, which involve the simulations of the model of the system being optimized. Many real-time control applications such as flight controllers, missile and plane trajectory optimization need new and fast computer algorithms. Parallel methods of optimization are not new. The approach described in this chapter is of this type, while objects are considered instead of physical processors. The gradient-based optimal control problems contain two main tasks which essentially influence the central processing unit (CPU) time consumption: 1. Evaluation of the gradient or conjugated search direction in the control space, 2. Evaluation of the object function. Simulation models have been used to compare the performance of the algorithm G and of a certain realization of the process Gy/z. The results can help to develop new fast parallel optimization algorithms for multiprocessor computers equipped with real-time parallel software tools.