This is a proceedings volume of a workshop on"Computationally intensive methods in Simulation andOptimization". The purpose of the meeting was to review andevaluate newly developed methods for combining simulationandoptimization. These methods are used for decisionproblems about stochastic systems like production systems orcommunication networks, which are too complex for beingmodeled in an analytic way. A simulation model for such asystem is the basis for an efficient use of recursivestochastic optimization techniques.The volume contains selected papers from three areas:1. Gradient techniques for discrete event simulation models2. Stochastic optimization methods3. Efficient random generation methodsThe book contains articles dealing with competing methodslike the score method, perturbation analysis andexperimental design techniques. For the first time, acomparision between these techniques is made.
I: Optimization of Simulated Systems.- Performance evaluation for the score function method in sensitivity analysis and stochastic optimization.- Experimental results for gradient estimation and optimization of a markov chain in steady-staten.- Optimization of stochastic discrete event dynamic systems.- Sensitivity analysis of simulation experiments: Regression analysis and statistical design.- II: Optimization and Stochastic Optimization.- A stochastic optimization approach for training the parameters in neural networks.- Integrated stochastic approximation program system.- Lexicographic duality in linear optimization.- Dual optimization of dynamic systems.- Stochastic approximation via averaging: The Polyak's approach revisited.- III: Random Numbers.- Nonuniform random numbers: A sensitivity analysis for transformation methods.- Nonlinear methods for pseudorandom number and vector generation.- Sampling from discrete and continuous distributions with c-rand.
Series: Lecture Notes in Economic and Mathematical Systems
Number Of Pages: 162
Publisher: Springer-Verlag Berlin and Heidelberg Gmbh & Co. Kg
Country of Publication: DE
Dimensions (cm): 24.41 x 16.99
Weight (kg): 0.3