Accession Number : ADA285951


Title :   Artificial Neural Network Metamodels of Stochastic Computer Simulations


Descriptive Note : Final rept. 25 Jun 1990-10 Aug 1994


Corporate Author : PITTSBURGH UNIV PA DEPT OF INDUSTRIAL ENGINEERING


Personal Author(s) : Kilmer, Robert A


Full Text : http://www.dtic.mil/dtic/tr/fulltext/u2/a285951.pdf


Report Date : 10 Aug 1994


Pagination or Media Count : 254


Abstract : A computer simulation model can be thought of as a relation that connects input parameters to output measures. Since these models can become computationally expensive in terms of processing time and/or memory requirements, there are many reasons why it would be beneficial to be able to approximate these models in a computationally expedient manner. This research examines the use of artificial neural networks (ANN), to develop a metamodel of computer simulations. The development and use of the Baseline ANN Metamodel Approach is provided and is shown to outperform traditional regression approaches. The results provide a solid foundation and methodological direction for developing ANN metamodels to perform complex tasks such as simulation optimization, sensitivity analysis, and simulation aggregation/reduction. Artificial Neural Networks, Computer Simulation Metamodel, Regression, Response Surface Methods, Simulation Optimization.


Descriptors :   *COMPUTERIZED SIMULATION , INPUT , OUTPUT , REQUIREMENTS , SIMULATION , NEURAL NETS , OPTIMIZATION , STOCHASTIC PROCESSES , MODELS , NETWORKS , COMPUTERS , PARAMETERS , PROCESSING , SOLIDS , SENSITIVITY , REDUCTION , TIME , REGRESSION ANALYSIS , SURFACES , RESPONSE , ARTIFICIAL INTELLIGENCE , APPROACH


Subject Categories : Computer Programming and Software


Distribution Statement : APPROVED FOR PUBLIC RELEASE