×

You are using an outdated browser Internet Explorer. It does not support some functions of the site.

Recommend that you install one of the following browsers: Firefox, Opera or Chrome.

Contacts:

+7 961 270-60-01
ivdon3@bk.ru

  • "Features of neural network modeling the inverse problem"

    The PLExp application which is a superstructure for Excel and Access on the basis of VBA is developed. The appendix is intended for automation of planning and carrying out multiple-factor experiments, turning on the following blocks: planning; carrying out experiment; regression and dispersive analysis; forecasting; assessment of errors. Application of the appendix allows to raise efficiency of the researches connected with carrying out natural or model experiments as gives an opportunity to reduce the number of the required experiments, to automate processes of storage, extraction, information processing and search of optimal solutions.

    Keywords: planning, factor, experiment, appendix, VBA, regression, dispersion, forecast, approximation error, function of a response, inquiry, modeling

  • Features of neural network modeling the inverse problem

    This paper investigates the possibility of applying neural networks to problems of reverse calculation of multi-layered road structure. The paper considers networks with different architecture. To improve the conditions of convergence are applied the back-propagation procedure with automatic selection step, different types of activation functions with varying parameters.

    Keywords: neural networks, inverse problems, polutorospalny predictor, the two-layer perceptron, the approximation of the activation function