CC BY
27
ALGORITHM OF INFORMATION SYSTEMS ADAPTIVE CONTROL UNDER THE CONDITIONS OF MASS PERTURBATIONS
BURENIN
Andrey Nikolaevich1 LEGKOV
Konstantin Evgenyevich2 ORKIN
Vadim Vitalyevich3
Information about authors:
1PhD, Docent, Chief Specialist of JSC «Research Institute "Rubin", Saint-Petersburg, Russia, [email protected]
2PhD, Head of the Department of Automated Control Systems of Military Space Academy, St. Petersburg, Russia, [email protected]
3Postgraduate Student of the Military Space Academy, St. Petersburg, Russia, [email protected]
ABSTRACT
The main feature of information systems as control objects is availability of both modern and prospective networks, software and hardware based on advanced technologies, as well as traditional networks with outdated equipment united by a single transport network for information transfer and by universal access procedure to transfer service. At some point in the information system they need to communicate with each other. In the conditions of increasing of the number, heterogeneity and complexity of equipment, the complexity of network and information technologies, of the network topology branching the task of managing the modern information system becomes one of the most important. The main task of control in this case is to ensure the functioning of an information system with specified performance indicators in conditions of mass perturbations. In general, the control process includes the following steps: obtaining information about the behavior of each information network and its equipment, analyzing the received information, developing and executing of the solution, i.e. control actions implementation on all objects of management (all networks, their control systems, their equipment). However, the introduction of centralized network control is difficult due to the heterogeneity of information technology and the types of equipment used in the network, including the network control systems themselves. This leads to a number of additional problems. The heterogeneity of network control systems can cause interruptions in information flows, slowing down the processes of generating control commands and increasing the probability of errors. One of the main tasks of the information system functioning control is the task of controlling the parameters that characterize the functioning of the system. This task can be solved on the basis of special procedures development for the formation of an information flow control plan (requests for the provision of services). The control plan is developed on the basis of the data distribution plans necessary for the work of officials at all nodes of the information system and determines the pre-determined order of the outgoing directions for the information transfer. The choice of this or that procedure for the formation of the control plan is recommended to carry out on the basis of the scientific and methodical apparatus of adaptive resource-saving information systems control, in which the adaptive control algorithm plays an important role.
KEYWORDS: algorithm; the order of choice; information system; flow of requests; control plan.
For citation: Burenin A. N., Legkov K. E., Orkin V. V. Algorithm of adaptive control of information systems under the conditions of mass perturbations. H&ES Research. 2017. Vol. 9. No. 6. Pp. 90-95.
INTRODUCTION
For the effective functioning of the currently created automated control systems (ACS) in complex organizational and technical objects it is necessary to provide timely information to managers and automation complexes with the required quality levels. For this purposes an information infrastructure control system with information systems at the heart is created.
Adaptive control appeared in the 50s of the XX century, when conventional control methods in complex technical systems could not be used. The questions of adaptive control are revealed in the works of Zemlyakov S. D., Rutkovsky V. Yu., Tsypkin Ya.Z. and Landau I. D. [1-2]. Adaptive management becomes relevant for information systems. The task of the information system functioning control includes control the parameters that characterize the information system functioning [4] based on the application of the procedures for the formation of the information flow management plan (requests for information). The problems of routing in communication networks, distributed automated systems and telecommunication networks were considered in the scientific works of a number of authors (M. Yu. Avanesov, A. N. Burenin, E. V Goncharov, G. P. Zakharov, A. P. Kuleshov, V. G. Lazarev, I. A. Mizin, N. Ya. Parshenkov, S. P. Prisyazhnyuk, S. I. Samoylenko, B. Ya. Sovetov, and others) [3-9]. However, a number of questions related to the possibility of using the developed methods and algorithms, the classification of which was carried out in [4, 10], in information systems. Their comparative analysis and determination of conditions for their most effective utilization remain open. Thus, it is necessary to consider the possibility of applying the procedures for managing the parameters that characterize the information system functioning under various conditions, including conditions of external influences on the information system. In the present work, a part of this tasks is solved by the development of the algorithm for adaptive control of information flows in the information subsystem of the ACS in complex organizational and technical objects, which is an actual scientific task and corresponds to the current trends in the development of information technologies.
FORMULATION OF THE PROBLEM
The task of controlling the parameters that characterize the information system (IS) functioning is a complex task and involves the utilization of procedures applied depending on the state of the IS.
As in the case of other control tasks of the IS, it is necessary in its functioning control that control goals be achieved. To do this, it is necessary to match the values of the efficiency indicators to required values in conditions of external influences.
In the functioning control plane it is advisable to set the efficiency of the IS with some functional F(t, L, Z, S, U U ), which depends on the fixed, unperturbed parameters and characteristics of the IS L, on disturbing external parameters Z,
on the state of the subsystem S, from the functioning control of the IS Uf on controlling the structure. S is the status of the information subsystem, S = (V, Y), where Vstands for the monitored state of the system (communication between nodes, the number of servers providing services, the degree of loading of nodes (servers)), Y is the uncontrolled state of the system (the intensity of receipt and maintenance of requests).
Considering F(t, L, Z, S, Uf U) as a function of functioning control, it can be argued that the choice of a specific procedure Uf = x will provide a very specific value of the IS efficiency indicator, where x is the decision to choose the procedure for controlling the operation of the IS. We represent the functional F in the following form fx, 5, z). As an effectiveness indicator of the IS, it is advisable to take a functional, depending on the size and connectivity of the IS structure, the performance and reliability of the IS components (state 5). Therefore, the solution of the task of controlling the functioning of the IS is to search for one U f (t) = x¡ : x¡ eA, where A = {x|p {/(x, s, z )> fr (x, s, z)}> P,^, with otfier things being equal, the performance indicator of the IS takes a value not less than the required one. Here is the "solution" or the state of the environment z e Z, 5 is the state of the system at the time of choice of x, m is the number of possible procedures for controlling the functioning.
ALGORITHM OF INFORMATION FLOW ADAPTIVE CONTROL
The previously developed methods of infocommunication systems and networks control [3, 4, 6] can be used for information systems control, with some limitations (the absence of a telecommunications component in the IS). IS can be in the conditions of the enemy's influence, so it should be about complex adaptive management, i. e. application of several options for changing the flow control plan, depending on the situation. In accordance with this, the algorithm of adaptive control of information flows in the IS is proposed in the paper.
The algorithm can be conditionally divided into parts that perform their "work" with a certain indicator of efficiency (the quality of service level for requests for the provision of information). At the initial stage functioning is performed based on the method of managing the processes of providing services with the formation of quality reliefs. Then, when information is received from the monitoring subsystem (its elements are located at each service node) that the quality of service level has decreased, the local adaptation procedures are connected, which operate on the basis of the received traffic data and queues to the service delivery servers.
In this case, the quality relief matrices continue to be corrected after a certain time Tf given before the entire system is operational. When the value of the performance indicator P > Prl is reached a return to functioning is made based on the method of managing the processes of providing services
with the formation of quality reliefs (to the normal mode of operation).
For P <Pr2, a transition to the functioning using global adaptation procedures (probabilistic-game procedures based on the statistics of the service of incoming requests). Formation of the management plan takes place with the help of a set of probabilistic vectors that collect statistics of successful (unsuccessful) attempts to transmit requests in various directions.
For P <Pr3 (significant structural changes and almost complete lack of information about the availability of servers on the nodes), a transition to the functioning based on the method of intelligent probe control. Here Ptr1, Ptr2, P tr3 are the values of the efficiency indicators determined before the system's operation (based on the results of simulation).
ALGORITHM FOR SELECTING THE DIRECTION OF INFORMATION TRANSFER BASED ON THE DISTRIBUTED PROGRAM CONTROL METHOD WITH GLOBAL AND LOCAL ADAPTATION PROCEDURES
The main part of the developed algorithm (Figure 1) is the method of distributed program control with global and local adaptation procedures (GLA), the essence of which is the formation of probabilistic matrices at distributed IS control centers that respond to the success of requests for information to those or other nodes of the IS.
The main advantage of all gaming procedures in general [7, 10], and in case of their application for IS parameters management in particular, lies in the fact that the formation of procedures and control plans does not require the transfer of any service information. Requests from officials and complexes of automation are service information. When creating a control plan for each new request, the results of their service in previous time are used, and after the maintenance this plan is corrected again.
Let the information system (IS) contain r nodes for the provision of services {SN1, SN2, ..., SNk, ... SN} and z nodes of the service consumers {N1, N2, ..., N1, ..., Nz}. Let SNk provide n types of information services. There is a stochastic matrix (2) in each /'-node of the IS—the recipient of the information service. In this matrix number of columns equal to r (the number of SNs to which the node can access (the number of outgoing directions)) and the number of lines equal to n (the number provided information services). Each j-th row of the matrix, corresponding to the j-th kind of services, is a vector:
A j = (P ji,P j 2,....,P jk,...,P jr ), (1)
where P 'ik — probability of access to the k-th service node),
k = Ü ; <Tpjt= 1.
П =
1 11 1 12 pi pi
21 22
pi1 pi
pi pi pi
1 n1 1 n2 1 t
(2)
Thus control plan matrixes are generated for each consumer node of the IS indicating the order of the choice of outgoing directions. When transmitting a request, the direction with the highest probability of transmission is selected.
This rule will be met if the probability of selecting the priority transmission direction is set to be proportional to the highest value of the quality of service indicator from the matrix of quality reliefs. As a given indicator we take the probability of timely service. For example, we apply to service 2 (second line of the plan), that
P 21 =
P =
22
P 2r =
02,
021 + 022 + ••• + 02 r
022
021 + 022 + • + 02 r
02 r
(3)
021 + 0
0
Accordingly, each IS node is associated with a matrix with a number of columns equal to r (the number of SNs to which the node can access) and the number of rows equal to n, which contains the values of the service quality indicators for each SN for each type of request, represented by the expression (4).
Q =
Ql1 Qi2
Q21 Q22
Q,1 Q„2
Q1r
Q2 r
Qnr
(4)
The outgoing direction y is selected. An action of the y-type is performed. If the request is served with the required value of quality indicator (information is provided to the consumer without exceeding the permissible time Td), then this direction is encouraged, otherwise it is fined.
An algorithm is presented that is an integral part of the algorithm of information flow adaptive control in the IS. The order of the algorithm: Step 1. The beginning of the algorithm. Step 2. Calculation of the probabilities of the k-th type action. Assignment of the vector A' The vector is given from the rule (1). The matrix (2) is formed by the formulas (3)
Step 3. Calculating Po — the average probability of requests servicing with quality below the required. Step 4. Calculation of a, p by the formulas:
k=1
a = -
P - fP
o Jo
P + fPo
O J O
p =
1 - Po + fPo 1 - Po - fPo
(5)
P _ PP
1+(P-1) P
Step 11. Comparison of m and value
P
(6)
the number
P =-
Pa
1 + (a- 1)P
(6)
Step 13-15. Elements of the corresponding line of the matrix are normalized after the penalty of direction (7) and after
the promotion (8) in such a way that V p = 1, j = 1, s.
¿—i jk J
P
P* _ jk jk ~ '
1 + (a- IP
P = jk
Pjk
1 + (P-1) P
, k ^ 3 ,k _ 1, r;
, k * 3 ,k = 1, r.
(7)
(8)
where f— coefficient obtained before the start of the functioning of the system (simulation) and showing the degree of influence of the results of service requests on the likelihood of subsequent selection of relevant directions.
Step 5. Receiving an application for the provision of a y-kind service. Variables Nap. serve to not select the same direction in retransmitting.
Step 6. In the y-th line of the matrix the maximum number is found among all P for k = 1, r, then the direction y is chosen equal to k, at which P is maximum to which corresponds to the number found. m = 1 (the first attempt to transmit in the current direction). k must not be equal to the already selected direction number.
Step 7. The action of type y (transmit in the direction of y) is performed.
Step 8. The action is performed
l = argmaxPjk,k y,k = 1,r,
k=1, r
where l is the second highest priority.
Step 9. Obtaining the result of an action of the type y. Step 10. If the application is served with the required quality, then the direction of y is encouraged:
of retry attempts required; m is the number of the attempt to transmit information (application) in the priority direction. If they are equal, the transmission attempts end, otherwise the transition to step 12.
Step 14. If not equal, then the number of the attempt is increased by 1 and the transition to the transfer of the request in the chosen direction.
Step 12. If the application was not served with the required quality, then the direction of y is penalized:
Step 16-17. Assigning new values to matrix elements
Pk = j 'k = ^ ■
Step 18. Checking the condition, if the request is served at all.
Step 19. If the application is not served, then if there is a need to apply local adaptation elements, a few more attempts are made to get service in this direction (priority) (go to step 8).
Step 20. If the application is served, then the conclusion "Request is served".
Step 21. If there is no need to apply the local adaptation elements, then assigning the auxiliary variable to the number of the current selected direction and transition to step 6. Nap is the number of the direction by which the transfer attempt was already made. If the flag about the application of the elements of local adaptation (LA) is present, then it is necessary to go to step 6.
Step 22. The end of the algorithm.
CONCLUSION
Along with the problems of integration of heterogeneous automated systems, there are tasks of improving of information interaction protocols. The algorithm of adaptive control of information flows in information systems is proposed. Suggestions and results of this work can form the basis for the operation protocols used in the program complexes for managing the information system of various control centers operating under perturbation conditions. In special purpose systems it is especially important to provide for the influence of external factors and impacts as a cause of the change in the state of the information system. Therefore, the integrated application of known methods for the distribution of flows in the IS control system is an urgent task for the newly created integrated information systems for special purposes, especially for distributed ones. Complex adaptive control algorithms can form the basis for management of information systems for special purposes.
REFERENCES
1. Zemlyakov S. D., Rutkovskiy V. Yu. Obobshchennyy algoritm adaptatsii odnogo klassa bespoiskovykh samonas-traivayushchikhsya sistem s model'yu [Generalized algorithm for adapting of the class of search-and-search self-tuning systems with a model]. Avtomatika. i telemekhanika. 1967. Vol. 28. No. 6. Pp. 88-94. (In Russian)
2. Tsypkin YA. Z. Adaptatsiya i obucheniye v avtomatich-eskikh sistemakh [Adaptation and training in automatic systems]. Moscow: Nauka, 1968. 400 p. (In Russian)
k=1
3. Burenin A. N. Ob upravlenii marshrutizatsiyey na os-nove modifitsirovannykh adaptivnykh metodov [On the management of routing based on modified adaptive methods]. Tekhn/ka sredstv svyaz/. 1991. № 7. Pp. 51-59. (In Russian)
4. Burenin A. N., Legkov K. E. Sovremennye /nfokommu-n/kats/onnye s/stemy / set/. Osnovy postroen/ya / upravlen/ya [Modern infocommunication systems and networks. Fundamentals of construction and management]. Moscow: Media Publisher Publ., 2015. 348 p. (In Russian)
5. Avanesov M.Yu., Prisyazhnyuk S. P. Operat/vnoye up-ravlen/ye potokam/ dannykh v mul't/serv/snykh setyakh svyaz/ [Operational management of data flows in multiservice communication networks]. St. Petersburg: Balt. state. tech. university, 2007. 81 p. (in Russian)
6. Lazarev V. G. Savvin N. G. Seti svyazi, upravlenie, kom-mutac/ya [Network communication, management, switching]. Moscow: Svyaz, 1973. 264 p. (In Russian)
7. Is'yanov V. M., Lazarev V. G., Parshenkov N. Ya. Det-sentral/zovannyy sposob d/nam/cheskogo raspredelen/ya /n-formats// na avtomat/chesk/ kommut/ruyemykh setyakh svyaz/ [Decentralized way of dynamic information distribution on automatically switched networks]. In book "Avtomaty i up-ravleniye setyami svyazi". Moscow: Nauka, 1971. Pp. 68-71. (In Russian)
8. Butrimenko A. V, Ginzburg S. L. Ob odnom sposobe det-sentralizovannogo dinamicheskogo raspredeleniya potokov in-formatsii [About one method of decentralized dynamic distribution of information flows]. In book "Information networks and commutation". Moscow: Nauka, 1968. Pp. 32-69. (In Russian)
9. Butrimenko A. V, Lazarev V. G. Sistema poiska op-timal'nykh putey peredachi soobshcheniy [The system of searching for optimal ways of transmitting messages]. Problems of Information Transmission. 1965. Vol. 1. No. 1. Pp. 80-87. (In Russian)
10. Novikov S. N. Klassifikatsiya metodov marshrutizatsii v mul'tiservisnykh setyakh svyazi [Classification of routing methods in multiservice communication networks]. Vestnik SibGUTI. 2013. No. 1. Pp. 57-67. (in Russian)
11. Gibbens R. J., Kelly F. P., Key P. B. Dynamic alternative routing — modeling and behavior. In Proceedings ofthe 12 International Teletraffic Congress, June 1988. Pp. 507-516.
12. Rainer Handel, Mafred N. Hunder, Stefan Schroder. ATM Networks. Concepts, Protocols, Applications. Addi-son-Wesley Publishing Company, 1999. 327 p.
13. Lindberger K. Analitical Methods for the Traffical Problems with StatisticalMultiplexing in ATM-Network. Tele-traffic and Datatraffic in a Period of Change. International Teletraffic Congress -13.Copenhagen, 1991. Pp. 807-813.
АЛГОРИТМ АДАПТИВНОГО УПРАВЛЕНИЯ ИНФОРМАЦИОННЫМИ СИСТЕМАМИ В УСЛОВИЯХ МАССОВЫХ ВОЗМУЩЕНИЙ
БУРЕНИН Андрей Николаевич, ОРКИН Вадим Витальевич,
г. Санкт-Петербург, Россия, [email protected] г. Санкт-Петербург, Россия, [email protected]
ЛЕГКОВ Константин Евгеньевич, КЛЮЧЕВЫЕ СЛОВА: алгоритм; очередность выбора; инг. Санкт-Петербург, Россия, [email protected] формационная система; поток заявок; план управления.
АННОТАЦИЯ
Основной особенностью информационных систем, как объектов управления, является наличие как современных перспективных сетей и программно-аппаратных средств, основанных на передовых технологиях, так и существующих традиционных сетей, объединенных единой транспортной сетью для передачи информации и универсальной процедурой доступа к транспортной услуге. Причем в рамках информационной системы им необходимо взаимодействовать друг с другом. В условиях увеличения количества, разнородности и сложности оборудования, усложнения сетевых и информационных технологий, раз-ветвленности топологии сетей задача управления современной информационной системой становится одной из важнейших. Основная задача управления при этом - это обеспечение функционирования информационной системы с заданными показателями эффективности в условиях массовых возмущений. В общем случае процесс управления включает следующие этапы: получение информации о поведении каждой информационной сети и ее оборудовании, анализ полученной информации, выработка и исполнение решения, т.е. осуществление управляющих воздействий на все объекты управления (все сети, их системы управления и оборудование). Вместе с тем внедрение централизованного сетевого управления затрудняется в связи с разнородностью информационных технологий и типов оборудования, применяемых в сетях, включая и сами системы сетевого управления. Это приводит к появлению ряда дополнительных проблем. Разнородность систем сетевого управления может вызвать
прерывания в информационных потоках, замедление процессов выработки управляющих команд и повышение вероятности возникновения ошибок. Одной из основных задач управления функционированием информационной системы является задача управления параметрами, характеризующими процессы функционирования системы. Данная задача может быть решена на основе разработки специальных процедур формирования плана управления потоками информации (заявок на предоставление услуг). План управления разрабатывается на основе данных планов распределения информации, необходимой для работы должностных лиц, на всех узлах информационной системы и определяет заданную на определённое время очерёдность выбора исходящих направлений передачи информации. Выбор той или иной процедуры формирования плана управления планируется осуществлять на основе научно-методического аппарата адаптивного ресурсосберегающего управления информационными системами, в котором важную роль играет алгоритм адаптивного управления.
СВЕДЕНИЯ ОБ АВТОРАХ:
Буренин А. Н., д.т.н., доцент, главный специалист АО «Научно-исследовательский институт «Рубин»; Легков К. Е., к.т.н., начальник кафедры автоматизированных систем управления Военно-космической академии имени А. Ф. Можайского;
Оркин В. В., адъюнкт кафедры Военно-космической академии имени А.Ф.Можайского.
ДЛЯ ЦИТИРОВАНИЯ: Буренин А. Н., Легкое К. Е., Оркин В. В. Алгоритм адаптивного управления информационными системами в условиях массовых возмущений // Наукоемкие технологии в космических исследованиях Земли. 2017. Т. 9. № 6. С. 90-95.
