CASPIAN JOURNAL
MANAGEMENT AND HIGH TECHNOLOGIES
ALGORITHM OF IMITATION CONTROLLED PROCESS IN A QUEUING SYSTEM
Article retracted 05.02.2018
 |
Read | Timofeev Paul A., Yagudaev Gennady G., Goryachkin Boris S., Stroganov Dmitry V. ALGORITHM OF IMITATION CONTROLLED PROCESS IN A QUEUING SYSTEM // Caspian journal : management and high technologies. — 2011. — №3. — pp. 62-70. |
Timofeev Paul A. - Ph.D., Advisor, Company "Centre for Analysis of Projects", 1/23 Staryi Petrovsko-Razumovsky pr-d, Moscow, 127287, Russia, kafedra@asu.madi.ru.
Yagudaev Gennady G. - Ph.D., North Caucasus branch of the Moscow Automobile and Road Technical State University (MADI), 64 Leningradsky av., Moscow, 125319, Russia, kafedra@asu.madi.ru.
Goryachkin Boris S. - Ph.D., Moscow State Technical University (MSTU), NE Bauman, 5 2-nd Baumanskaya str., Moscow, 105005, Russia, kafedra@asu.madi.ru.
Stroganov Dmitry V. - Ph.D., Moscow State Technical University (MSTU), NE Bauman, 5 2-nd Baumanskaya str., Moscow, 105005, Russia, kafedra@asu.madi.ru.
The article investigates the methods of using simulation models to solve the problems of parametric synthesis, which leads to the choice of values of variable parameters, delivering extreme objective function. The range of values of variable parameters is determined in accordance with the possibilities of parameterization of the model. It is assumed that computes the objective function using a simulation model of the object, which allows us to calculate the objective function for all values of variable parameters in a given range of parameters. For each value model unambiguously defines the value of the objective function. Since the procedure of simulation and optimization algorithms require substantial computational resources, the article suggests to combine the processes of search engine optimization and simulation models in a single manageable. Completed formal description of a controlled simulation model and highlights the main process in the space of the resulting performance and process management in the space of control parameters. We propose a control algorithm of gradient type, belonging to the class of stochastic approximation algorithms are defined scheme of its implementation and the basic parameters. The analysis of the convergence of the algorithm for different plans: a central, symmetric, and the plan with a central point. It is shown that the plan with the central point has the best characteristics. Variance of estimate are close to it by the symmetric and central plans, and differ slightly from each other. However, the central plan requires half the number of measurements to determine the assessment of the extremum, which is important for simulation. The results showed robust to the initial state and the variance of the convergence of the algorithm with the symmetric and the central plan.
Key words: simulation,Gaussian processes,mean integral evaluation,variance,trend,modeling,non-stationary processes,autocorrelation,convergence,transition regime,queuing system,recursive analysis.
Причина ретракции: материал статьи в основном заимствован из докторской диссертации В.М. Черненького «Процессно-ориентированная концепция системного моделирования АСУ» (защищена в 2010 г. в г. Москве).
Решение о ретракции принято редакционной коллегией журнала «Прикаспийский журнал: управление и высокие технологии» 05.02.2018 г.