Adaptive Control of Stochastic Integral Calculating Processes

摘要

The adaptive control theory of stochasitc systems is applied to quite artiicial objects, such as Monte-Carlo computation procedures of many-dimensional integration. The control action here is represented with function of unit distribution density of random grid of integration. The criterion of evaluation accuracy is chosen to be the criterion of optimal functioning, since it has analytical representation for both considered objects of control. The purpose of the paper is to apply the methods of the adaptive control theory in adjusting and adapting calculative algrithms while performign them. Such an approach appears to be effective due to that it enables obtaining analytical solutions for variational problems of optimization of accuracy of distribution density of integration grid knots in the algorithm. The methods of adaptive control being applied provide significant incease of computing processes efficiency.