CC BY
29
cptrPtrl = strchr(cptrFirst, if(cptrPtr1 == NULL) cptrPtrl = cptrPtrEnd;
else
cptrPtr1++;
// Working with the list of blocks: pList01 = ptrBlockList; if(pList01 != NULL)
pList02 = pList01->pNext;
else
pList02 = ptrBlockList; while(pList02 != NULL) { pList01 = pList02; pList02 = pList01->pNext; } pList01->pNext = new BLOCK_LIST; pList01->pNext->pNext = NULL; pList->JmpType = TypeOfJump(carrJumps); pList->pBegin = cptrPtr1; pList->pEnd = (cptrPtrEnd - cptrPtr1); } return TRUE;}
6. Заключение
Для анализа оптимальной (рациональной) структуры MISC-процессора были выполнены такие исследования:
— рассмотрены варианты сегментации глобальной микропрограммы в целях локализации ошибок при анализе и выборе оптимальной структуры ММП;
UDC658.012.011.56 ‘
HYBRID LOGIC-DYNAMIC MODEL FOR PROBLEMS OF THE SITUATIONAL CONTROL OF THE COMPLEX TRANSPORT SYSTEMS
KAMAL AHMED S.K., KARGIN A.A., RAHMAN A TIQ UR, SITNIK B. T._________________________
In the present work, for the class of the transport-problems related with the logic-programmed control, the integration of the two classes of models of control - situational control in real time and the logic- dynamic control where the change of the situation takes place through the model of the production situation in the modeling time is proposed. The model was successfully tested in the city transport network of Kharkov.
1. Introduction
For the control of the complex systems the situational model is getting wide spread use [1,2]. Being formed on the basis of the theory of control for the large systems, the situational control is oriented to the use of the knowledge about the object and the methods for its control. The mediums for the formal representation of the situation in the form of the language of the situational
— получены правила определения по внешнему виду выделенного фрагмента конкретной работоспособной структуры ММП из набора имеющихся структур;
— произведена оценка каждой структуры, исходя из характеристик данных аппаратных средств.
Литература: 1. Новосёлов В.В. Метод хронооптимизации блок-схем алгоритмов микропрограммного управления // Автоматика и вычислительная техника, 1983. №3 С. 25-28. 2. Новосёлов В.В., Шумилов Л.А. Выбор структуры микропроцессора на комплекте микропрограммируемых БИС // Управляющие системы и машины, 1983. №3. С. 21-24. 3.НовосёловВ.В., ПарфененокВ.Л. Оптимизация системы синхронизации микропроцессора методами математического программирования. // Деп. ВИНИТИ от 23.12.82, № 6345 — 82 ДЕП. 31 с. 4. Бережная М.А. Лобода
B. Г., Цуканов В.Ю. К вопросу проектирования структуры процессора // Радиоэлектроника и информатика, 1998, №2(3) .С.120-124. 5. Комплект БИСК1804 в процессорах и контроллерах / В.М. Мещеряков, И.Е. Лобов,
C. С. Глебов и др. М.: Радио и связь, 1990. 256 с. 6. Сизов К.А. Микропроцессор будущего: RISC, CISC или MISC?/ / Библиотека информационных технологий: Сборник статей. Вып.1 / Под ред. Г.Р. Громова. М.: Наука, 1990. С.118-124. 7. Ельчанинов Д.Б.,. Лобода В.Г., Цуканов В.Ю. Модели архитектуры MISC - процессора// Радиоэлектроника и информатика, 1999, №1(06). С. 85-89.8. Язык программирования Си / Бриан В. Керниган, Деннис М. Ритчи. М.: Финансы и статистика, 1992. 270 с.
Поступила в редколлегию 17.01.2000
Рецензент:
Цуканов Виталий Юрьевич, аспирант кафедры АПВТ ХТУРЭ. Научные интересы: алгоритмическое и программное обеспечение функционально-ориентированных процессоров. Адрес: 61111, Украина, Харьков, ул. Познанская, 2, кв.67,тел 10-42-63.
control [1,2] ofthe discrete situational network [3], RX-codes and frames 111 are related to the semiotic model for control [4] for which there exists the problem of automatic formation of the model of the situation.
The above mentioned problem can be solved by using the fuzzy models [5,7] if the following two problems are solved: 1)to work out the automatic formation of the structural model of the production-situation on the basis of the information received in the real time mode from the control-measurement instruments and sensors; 2)to introduce the time factor in the fuzzy model of the situational control.
The first problem is solved through introduction of the model of the situation which reflects the production-situation expressed through structured multilevel fuzzy set [8-10] at every moment ofthe real time. The membership function of the elements of the lower level is formed automatically on the basis of the data received from the sensors and the control-measurement instruments and the membership function of the upper level reflects the structured knowledge about the object of control and its environment formed automatically according to the declarative rules. In this way, at every moment of time, an automatically formed fuzzy set represents the model of the real situation. The technology for the formation of the models of the hybrid (accurate and fuzzy) structured situation is given in [10-11].
РИ, 2000, № 1
63
The second problem can be solved by two methods: to introduce the clear time-dependence ofthe membership function of the fuzzy set on the basis of which the model of the situation is formed and also the default time-dependence. The first method is based on the basis of the discrete control models - finite automata, neural networks 1\11, system of the impulse regulation [13]. For them it is specific that the time is divided into equidistant intervals and it is assumed that the change in the parameters occurs in the discrete moments of time. The clear dependence of the membership function on time in the situational models is caused firstly, due to the sharp increase in the number of rules in the knowledge-base and secondly, since each of the rules reflects an image ofthe fragment ofthe situation, hence there arises the necessity ofthe determination ofthe moment of time when a rule will be applied to the models ofthe current situation. The latter is related either with the extreme difficulty in testing all the rules at each moment oftime or with the danger of escaping the optimal moment of applying the controlling decision.
Hence for the class of the transport-problems related with the logic-programmed control the default method is applied. The time factor unclearly exists in the models representing the current situation and in the rules for taking the control-decisions, but clearly exists in the models representing the change ofthe situation with time [8-10]. Model ofthe situation at every moment ofthe real time reflects the condition of the object of control and its environment. Controlling rules are applied to the model of situation only at the moments of time when the value of the membership function of any of the elements changes with a value greater than some given value e. Hence we get an uneven accidental series ofthe moments oftime of formation ofthe model of situation and control. This allows considering the rules for the formation of the models of situation (declarative) and the rules for the formation of control (production) as statical in spite ofthe fact that the control is realized for the dynamic objects.
The given class of model can be used also for solving the problems of control which requires to consider the dynamic parameters of the object of control, e.g. instant speed of change or the integrated characteristics of the parameters. Due to this reason the basic definition of the fuzzy set is changed as
s = {Si I Si) I ^(Si)}*! , (1)
where S is the fragment ofthe situation representing the dynamic fuzzy set; si is an element of the set; p(si) is
the membership function of the element Sj to the set S;
ra(si) is the instant speed of change of the membership
function p(si) at the current moment of time.
The dynamic fuzzy set (1) is used for the formation of the system of the situational control with the speed of release of the trains as component objects on the railway-distributive hills in different directions.
2. Formal Representation
The integration of the two classes of models of control -situational control in real time and the logic- dynamic control where the change of the situation takes place
through the model of the production situation in the modeling time. In the picture is shown the structure of the knowledge base according to which this integration is realized.
Structural scheme for Hybrid Logic-Dynamic model
At the moment of taking decision about discharge control according to the rules of BPK_l, the analysis of the production situation is performed. If in it there exist the fragments satisfying the conditions for the global discharge control, then starts the function of the logic-dynamic model (LDM). At the first stage the model of the generalized situation is formed in which the LDM works. In fact, these are the parameters of the LDM, e.g. initial distribution of the moving objects (MO) along the routs, which correspond to their real condition of the problem of the discharge control for movement. Model of the situation LDM is formed according to the declarative rules from BPK_1. LDM realizes the modeling of the change ofthe situation, e.g. movement ofthe MO in the modeling time modifying at every step the model ofthe generalized situation, i.e. the results of the modeling are reflected in the model of the situation in which the LDM works.
At every step the model of the generalized situation is analyzed by the rules from BPK_2 and BPK_3. The evaluation rule ofthe generalized model BPK_3 tests the situation ofthe object through the presence ofthe fragment of the goal, at which the modeling stops and the control decision is taken. Rules for the modification of the parameters LDM (BPK-2) analyze the model of the generalized situation ofthe object for the existence of the fragments which represent “bad” in the produced situation and causes the changes in the parameters ofthe LDM after which starts the modeling from the initial situation.
At the moment when the modeling is finished which is determined by the rules from BPK_3, there form the control-decisions, which are taken to the model of the produced situation in real time. This is performed according to the rules from the DBK_2. The class of problems for controlling the transport related MO which is studied in the present work is related to the situation
64
РИ, 2000, № 1
containing the following portion of information entering in the intellectual control system (ICS) at the discrete moment of time when the MO passes through the controlling point: 1) number of the controlling point; 2) passing time through the controlling point; 3) number of the MO; 4) number of the routs.
3. Logic-Dynamic Model
Since the size of the net model is the least, it is chosen as the base model. Applying to the city electrical transport (CET) trams this model contains 55 knots which represent the changing points of routs and the ending points and also 66 arms each of which represent the connecting rail Unes between two knots. Each group CW (Z) can be presented as the product of the periodically
"k.j
repeating distribution of the zero MO on the k-th rout
1
c^w(Z)-
1-Z
-Mk
Mk
t C?ZfUrV.
v=0
(2)
In the places of all the jє (0,N) MO CknkXjj(z) on this rout
Nk
C^JJ(Z) = sz%kjj
j=o ’
Vk = V(0,l,2,...,Mk),
1 Nk
----^zr^lz^j-'
j=o ’
i.e. cLj(Z) =
Z M v=0
operator for determining the integral part;
integral part from
Let us
present two zero distribution Ck. j yr - ) in the form of the
I
series of fractions of the type {r _ , the integral part of
which shows the number of the knot of the net, nominator of the fraction shows the time ZvvT for passing the MO through this knot in the straight direction, and the denominator shows the time Zv vT for passing the knot of the net in the opposite direction:
C
1.0 \
= (7 — ,6 — ,8 — ,16—,17 — ,18 — ,19 — \ 21 18 16 15 13 12 11
(3)
C
29.0\
0 2 12 19
54—,49 —,51 —,52— 39 37 27 20
With the help of (3) let us write the reflection of C}' 0(Z)
for the distribution of the zero MO of the first rout in the straight and opposite directions
Ci oh) as Ci.o(Z) = jci.o(h
еда— szr+zzh =
^ 1 v = n v = l 1
(4)
Let us consider that the MO of the first rout is placed serially one after the other at an interval of pjZT where Pj = 2 . In this case all the N1 groups of the first rout can be presented through the following model for
:j-Ml
N,=FKW’ci j(Z) =
N,=FIXy) = ,0.
(5)
The found model allows at any moment of time (i = ц+ j ) determine the place of existence of any MO in the set (i, г, M, v, v, j,k, pk) as the distribution of the projection of the group Ck j on the i-th axis of time:
CL.J = (CL.O.CL.,...CL.
Nk
(6)
РИ, 2000, № 1
65
4. Declarative Knowledge Base - Rule Of Formation Of The Generalized Model Of Situation
The problem which is solved with the help of the rule BPK_1 is to form the distribution ofthe MO on each rout, i.e. to fix the parameters pj of the model (5), where pj is the distance in the unit of the interval of modeling time of the l-th rout of the MO with respect to the (l-l)-th rout. In this way, the fragment ofthe situation-structure LDM representing the initial placement of the MO can be presented in the form
C = {C1,C2,...,Ck,...,C30},
(7)
here Ck is the fragment of the structure of the initial placement of the MO on the k-th rout
C = {C1,C2,...,Ck,...,C30},
(8)
where 7j - {11 дО)Д І /И(2)>3 I,...,25 І /л{25)} — is the initial set the elements of which correspond to the interval
p^. For example, ifthe second MO on the k-th rout exists at a distance of 5 units of the modeling time with respect to the first MO, then we get
T74 = {1|0,2| 0,3 I 0,41 0,5 11,61 0,7 | 0,...,25 | 0}.
Formation ofthe structures (7) and (8) are realized on the basis ofthe fragment ofthe situation “placement of MO”.
Rules For The Formation Of Control
Result of modeling the movement of the MO along the entire network of the routs is ^ which is present in (4) and (5). The given values are needed to be reflected in the recommended speed of movements of the MO on the corresponding sections of the routs. As it was mentioned earlier that the recommended speed is applied on the MO in the form of three elemental evolved fuzzy set
<<SPEED={small| //, pj, medium| p2, big| p3 }>>.
The given correspondence is established with the help of the rules from DBK_2. For this the formed fuzzy set ^
with which the LDM works, is needed to reflect in the fuzzy set <<SPEED={small,medium,big}>> as the
elements of the set << % = {A^, A,™, A,3} >>. In that case
the mechanism of reflection will be the trivial one and the rule from the DBK_2 takes the following form
small med big
Xi 1.0 0.0 0.0
where IF01 — A2 0.0 1.0 0.0
A3 0.0 0.0 1.0
6. Conclusion
The proposed model allows consider the situation at any moment of real time and take the decision. The model being applied to the city tram-network allows at moment of time determine the place of existence of different trams in different routs. The model was applied in the transport control system of Kharkov city and was found to be effective.
Bibliography: 1.Крылов Ю.И. Ситуационное управление большими системами. М.: Энергия,1974. 134р. 2. Поспелов ДА. Ситуационное управление: теория и практика. М.:Наука. і986. 288р. 3. Железов Ж.И., Поспелов Д.А. Дискретните ситуационни мрежи — подкласс на големи-ти системы //Автоматика и вычислительная техника, 1969. № 2. С. 5-12. 4. Осипов Т.С. Две задачи теории семиотических модлей управления. III Представление семиотических моделей // Изв.АН СССР. Техническая кибернетика, 1981. № 5. С. 97-109. 5. Мелихов А.Н., Берштейн Л. С., Коровин С.А. Ситуационные советующие системы с нечеткой логикой. М: Наука, 1990. 272 с. 6. Алиев Р.А., Церковный А.Э., Мамедова Г.А. Управление производством при нечеткой исходной информации. М: Энергоатомиздат. 1991. 263 с. 7. Аверкин
A. Н., Батырин Н.Э., Билук А.Ф. и др. Нечеткие множества в моделях управления и искуссвенного интеллекта. М: Наука, 1991. 76 с. 8. Каргин А.А., Демин В.А., Новиков
B. Б. Ситуационная продукционная система управления технологическими процессами в производстве нанесения гальванопокрытий (СПРУТ-1)//Приборы и системы управления, 1991. №3. С. 6-8. 9. Каргин А. А., Петренко Т.Г. Статические модели представления и обработки ситуаций в системах ситуационного управления реального времени / Искусственный интеллект, 1999. № 1.
C. 35-41. 10. Каргин А.А., Петренко Т.Г. Представление и обработка декларативных знаний в ситуационных инте-лектуальных машинах /Наукові праці Донецького державного технічного утніверситету. Серія «Інформатика, кібернетика та обчислювальна техніка». Вип. 6: Донецьк: ДонДТУ. 1999. С.315-321. 11. Каргин А.А. Об использовании нечетких знаний в задачах управления движениям поездов. Ч. 1. Проекционные модели знаний на основе нечетких множеств // Информационно-уп-равляющие системы на железнодорожном транспорте. Харьков: ХАРАЖТ. 1996. № 6. С. 25-27. 12.Каргин А.А., Сытник Б. Т. Об использовании нечетких знаний в задачах управления движением поездов. Ч. 2. Структурированные, декларативные и процедурные знания в продукционных системах / Информацонно-управляющие системы на железнодорожном транспорте. 1997. №1. С. 4044. 13. Сытник Б.Т. Теоретические основы моделирования дискретно-динамических транспортных систем // ИУСЖТ, 1998. №4. С.42-47.
Поступила в редколлегию 02.12.2000
Рецензент: д-р. техн наук Скобцов Ю.А.
Kamal Ahmed S.K. Ph.D. in engineering science, Associated Professor, Department of Applied Physics and Electronics, Rjshahi University, Rajshahi; Scientific interest: Knowledge engineering; Hobby: fishing; Address: Department of Applied Physics and Electronics, Rjshahi University, Rajshahi, Bangladesh. Tel: ++88 - 0721 - 750041.
Kargin A.A. D.Sc. in engineering science, Professor, Department of Computer Technology, Donetsk State University; Scientific interest: Knowledge engineering; Hobby: ; E-mail: [email protected] Tel.(062)337-30-74.
Rahman Atiqur, M.Sc. in computer science, Research worker, Department of Applied Physics and Electronics, Rjshahi University, Rajshahi; Scientific interest: Artificial Intelligence; Hobby: driving; Address: Department of Applied Physics and Electronics, Rjshahi University, Rajshahi, Bangladesh. Tel: ++88 - 0721 - 750041.
Sitmk B.T., Ph.D. in engineering science, Associated Professor, Department of Automatic & Computing Control system Kharkov’s Academy of Railway Transport. Scientific interest: Knowledge engineering. Tel. (0572) 20-60-59.
66
РИ, 2000, № 1
