RUS | UKR || Master's portal of DonNTU
Reutskaja Natalia

Reutskaja Natalia

Faculty: Computer Science

Speciality: Economical cybernetics

Theme of master's work:

Analysis of offectiveness of grocery store functioning by Markovian models practice

Scientific adviser: Feldman Lev Petrovich


About author

Introduction

In the operation research we have to come across with the work analysis of the peculiar systems called the systems of mass service (SMS). For an example they are telephone systems, repairing workshops, ticket offices, information offices, shops, hairdressers and others. Each SMS consists of some servicing units which we will call servicing canals. In case of “canals” may figure connection lines, work places, devices, railways, lifts, cars and others. The systems of mass service may be one-canal and multi- canal ones. Each SMS is intended for the servicing of an application flow (or “demands”) received on SMS at any occasional time. The servicing of received application lasts some occasional time, and then the canal is free and ready to receive to get another one. The occasional character of application flow leads to the accumulation of the large application quantity on the SMS entry at some periods of time. These applications form a queue or they leave the SMS unserved. At other periods of time the SMS will work with under loading or idle at all. Each system of mass service, depended on the canal quantity and their productivity with the character of application flow, has a blotting capability which allows it to manage with application flow more or less successfully. The subject of mass service theory is the definition of dependence between the character of application flow, canal number, and their productivity, SMS work rules and service effectiveness. For the description of a food shop function the SMS with time limit expectation suits more. Let’s consider such SMS staying in markovskaja scheme. Let’s suppose that there is a canal SMS with expectation in which the quantity of places at the queue is not limited but the time in the application queue is limited by the occasional term T with the average meaning T . So each application is influenced by “leaving flow” with the intensity

(2.1)

If this flow is puassonovskij so the process in SMS will be markovskij. Let’s find the condition probability for it. Let’s numerate again the system condition to application quantity connected with the system as serviced and those in the queue.
  • S0 —All canals are free,
  • S1— One canal is busy, the rest canals are free,
  • …………………………………………………
  • Sh — K canals are busy, the rest canals are free,
  • …………………………………………………
  • Sn — All n canals are busy,
  • Sn+1—All n canals are busy, one application is in the queue,
  • ………………………………………………………………..
  • Sn+r — All n canals are busy, r applications are in the queue,
  • ………………………………………………………………
  • Sn+m - All n canals are busy, m applications a in the queue and so on and so forth.

    Condition diagram is shown on Fig. 2.1

    Let’s consider this diagram and put near the pointers appropriate intensities. As usual all left right pointers will have intensity of application flow. For the condition without queue which is from their right to left the sum intensity of servicing flow with all busy canals will be. To deal with queue condition the pointers from right to left the sum intensity of servicing flow of all n canals nm plus the appropriate flow intensity with leaving queue will be. If there are r applications the sum flow intensity of leaving will be rv. As we can see from the diagram there is a scheme of destruction and reproduction before us: using general meanings for limited condition probabilities in this scheme we will write down:……..

    Or taking signs.

    (2.3) Relative blotting capability q of such SMS may be counted as following. All the application are certain serviced except those which leave the queue in advance. Let’s count which average number of application leave the queue in advance. For this let’s count the average application number:

    r = 1* pn+1 + 2*pn+2+…+r*pn+r+… .

    Each application is influenced by “leaving flow” with intensity v. it means that from average application number r in the queue on average vr applications at a time unit will leave unserved: at a time unit on average will be served

    applications. Relative blotting capability of SMS will be

    2.4

    The average number of busy canals z will be received dividing absolute blotting capability into m:

    2.5 It allows counting the average number of applications in queue г not summing the unlimited row (2.2) really we will get from 2.5

    2.6

    We can find the average number of busy canals included into this formula as mathematic expectation of occasional value Z which has the meaning 0, 1, 2, …., n with probabilities…

    Z= 0*p0+ 1*p1+2*p2+…+ n*[1-(p0 +p1 +…+pn-1)]=p1+2*p2+…+ n*[1-(p0 +p1 +…+pn-1)] 2.7

    Note that in formula (2.1) the sum of unlimited row is figured which is not progression. But this sum is counted approximately and easy enough as the row parts decrease quickly with their number increasing. As the approximate meaning for unlimited sum we will take the sum of the final figure г – 1parts but the rest is estimated as following:

    2.8

    In the conclusion we note that if in the formulas (2.1) we pass to the limit v - 0 (b- 0) in case ….p n the formulas (6.10) are received of the previous paragraph, so “impatient applications” will became “patient”.

    The conclusions.

    We come across with the necessity of being in expectation condition in all person practical activities. Such situations occur in queues to ticket offices, big airports, on the telephone stations in case if the line is engaged, in repairing workshops in case expecting lathe and equipment repairing and in supply-sale organizations in unloading or loading of transport means expectation. In all these cases we deal with mass servicing. The mass servicing theory is involved in such situations studying.

    The occasional character of application flow, but in general case and the servicing duration, leads to the occasional process in the system of mass servicing.

    It is necessary to study the occasional process in the system, to describe it mathematically to give the recommendation in the rational organization of this process and to make sensible demands to SMS. The mass servicing theory deals with such problems.

    Literature

    1. Вентцель Е.С. «Исследование операций» - М.:Советское радио, 1972 ,552 стр
      URL http://www.4tivo.com/education/4031-e.s.ventcel.-issledovanie-operacijj..html
    2. Ивченко Г.И.,Каштанов В.А.,Коваленко И.Н.«Теория массового обслуживания:Учеб. пособие для ВУЗов» 1984, 257 стр
      URL http://dic.academic.ru/dic.nsf/ruwiki/214892
    3. Е.Крылов «Моделирование систем массового обслуживания конференция «Люди в экономике»06.04.2005: Что такое система массового обслуживания, примеры и классификация.»
      URL http://www.real-time-enterprise.ru/staff/article/queue/qs11.html
    4. Ослин Б.Г. «Технология имитационного моделирования систем массового обслуживания». Первая Всероссийская научно-практическая конференция «Опыт практического применения языков и программных систем имитационного моделирования в промышленности и прикладных разработках" ИММОД 2003
      URL: http://www.gpss.ru/immod'03/027.html
    5. Юдинцев А.Ю. и Ошкало В.В. «Информационные технологии и проблемы современного менеджера» Алт. ун-та. Барнаул 2004
      URL: http://ec.asu.ru/lib/sborn/regec2003/pdf/26.pdf
    6. А. Хэмди Таха «Введение в исследование операций», Вильямс, К-2001, 913с
      URL: http://financepro.ru/management/5544-takha-khemdi-a.-vvedenie-v-issledovanie-operacijj.html
    7. Н.И. Щедрин, А.Н. Кархов. «Экономико-математические методы в торговле» Экономика, М, 1980. - 176 стр.
    8. А.Я. Хинчин «Математические методы в теории массового обслуживания», М, 1955.-420 стр.
    9. Ю.К. Беляева «Теория массового обслуживания», М,1972. 168 стр.
    10. А.А. Спирин, Г.П. Фомин. «Экономико-математические методы и модели в торговле», Экономика, М, 1988.-192 стр.