|
ИСТИНА |
Войти в систему Регистрация |
ИСТИНА ИНХС РАН |
||
STATISTICAL ESTIMATIONS OF TRANSACTION FLOWS IN MULTI-COMPONENT INFORMATION SYSTEMSстатья
- Авторы: Larkin E.V., Bogomolov A.V., Ivutin A.N.
- Сборник: Распределенные компьютерные и телекоммуникационные сети: управление, вычисление, связь (DCCN-2017)
- Год издания: 2017
- Первая страница: 596
- Последняя страница: 603
- Аннотация: Transaction flows in multi-component information systems, which functioning may be represented as generation and servicing of quests, are investigated. It is shown, that analytical simulation of such systems may be substantially simplified, when transaction flows are the Poisson ones. The problem of statistical estimation the degree of approach of transaction flows to the Poisson flow is solved. The Pearson's and parametrical criteria are investigated and computational complexity of statistical estimation of intervals between transactions by such criteria is evaluated. New criterion, based on estimation of waiting function expectation is obtained. It is shown that use of such criterion permits to obtain the degree of approximation the properties of flow under investigation to Poisson flow, with minimal spending of computer time. Eligibility of proposed statistical estimations is verified with use the Monte-Carlo method.
- Добавил в систему: Богомолов Алексей Валерьевич