Publication: Kayıp Müşterili Kuyruk Modellerinin Analizi Ve Bir Uygulama
Loading...
Date
Authors
Journal Title
Journal ISSN
Volume Title
Abstract
Ill KAYIP MÜŞTERİLİ KUYRUK MODELLERİNİN ANALİZİ VE BİR UYGULAMA ÖZET Bu çalışmanın amacı GI I M l\ln~\ stokastik servis sisteminin temel göstergesi sayılan müşterinin kaybolma olasılığını incelemek ve bu sonuçları Samsun borsa işletmesinde uygulamaktır. Bu nedenle sonlu kapasiteli kuyruk modelleri araştırılmış ve konu ile ilgili literatür geniş çaplı olarak giriş kısmında belirtilmiştir. Çalışmamızda Gl İM l\ln-\ kuyruk modelini temsil eden yarı-Markov süreci tanımlanmıştır. Daha sonra tanımlanan yarı-Markov süreci yardımıyla müşterinin kaybolma olasılığı için pratik bakımdan kolay olan yeni bir formül türetilmiştir. Bu formül ak bir hizmet süresi içerisinde sisteme k-tane müşterinin gelme olasılığı olmak üzere a0,a,,...,a'_, elemanlarından oluşan özel bir determinantla verilebilir. Buna ek olarak elde edilen formül kullanılarak Gl I M İM n-\ kuyruk modelinin özel durumları olan sonlu kapasiteli kuyruk modellerinde müşterinin kayıp olma olasılığı hesaplanmıştır. Ayrıca ortalama ilk kaybolma anı ve ardışık kaybolma anları arasında geçen ortalama süre yine aü,a^,...,an_l elemanlarından oluşan determinantla hesaplanmıştır. Çalışmada ayrıca kaybolan müşteri akımının asimptotik analizi yapılmış ve a hizmet süresinin gelişler arası süreden büyük olması olasılığı olmak üzere a - > 0 koşulu altında kaybolan müşteri akımının Poisson akımına yaklaştığı gösterilmiştir. Son olarak gelişler arası sürenin sabit olduğu durumda G//M/1/0 kuyruk modelinde müşterinin kaybolma olasılığının minimum olduğu gösterilmiştir. Çalışmanın uygulama kısmında ise incelenen kuyruk modelinin özel durumu olan bir hizmet kanallı ve 3 bekleme kapasiteli GI I M 1113 modeline uyan Samsun borsa işletmesinden alman veriler değerlendirilmiştir. Elde edilen formüller yardımıyla Samsun borsa işletmesinde müşterinin kaybolma olasılığı, ortalama ilk kaybolma anı ve ardışık kaybolma anları arasındaki sürelerin ortalaması hesaplanmıştır. Anahtar Kelimeler: Sonlu kapasiteli kuyruk modeli, müşterinin kaybolma olasılığı, yarı- Markov süreci
IV ABSTRACT The aim of this study is to analyse the probability of loss of customers as a fundamental sign of GI / M 1/n-l stochastic service system, and to implement the system in Samsun stock business administrotron. For this reason queue models with finite capacity have been investigated, and related literature has been given in detail in the first chapter. In materyal and method, some matematical concepts in the study have been presented and it has been clasified how these concepts have been used in the study. In Findigs and Discussion, semi-Markov process, which represents GI / M 1/n-l queue model has been defined. Then a new formula which can be practically easy for the probability of customer loss by the help semi-Markov process has been found. This formula can be given by a specific determinant which consists of the aQ,al,...,aivA elements when there is a probability of customers loos in number k in the system within ak sevice period. Additionally, the probability of custemer loss in queue models with finite capacity as spectral cases for GI / M 1/n-l queue models have been calculated by the formula found. Besides, the avarage interval between the mean first loss moment has been calculated by the determinant of a{),al,...,an_[ elements. The asimptotic analysis of the lost customer stream has beeen also done, and it has been verified that the customer stream, which disappears in a - » 0 approaches to Poisson stream when a service period is longer than the inter-arrival time. Last in that chapter it has been also proved that the probability of customer loss in the GI / M 1/0 queue model has been minimum if the inter-arrival time is kept fixed. In the chapter of practice, the findings taken from Samsun stock business administration in accordance with GI / M 1/3 model with three waiting capacity and service variant that is a spectral case of the queue model in question have been evaluated. The probability of the customer loss in Samsun stock business administration, and the avarage interval between mean first loss moment and adjacent loss moment have been calculated by the help of the formula found. Keywords: The queue model with finite capacity, the probability of customer loss, semi-Markov process
IV ABSTRACT The aim of this study is to analyse the probability of loss of customers as a fundamental sign of GI / M 1/n-l stochastic service system, and to implement the system in Samsun stock business administrotron. For this reason queue models with finite capacity have been investigated, and related literature has been given in detail in the first chapter. In materyal and method, some matematical concepts in the study have been presented and it has been clasified how these concepts have been used in the study. In Findigs and Discussion, semi-Markov process, which represents GI / M 1/n-l queue model has been defined. Then a new formula which can be practically easy for the probability of customer loss by the help semi-Markov process has been found. This formula can be given by a specific determinant which consists of the aQ,al,...,aivA elements when there is a probability of customers loos in number k in the system within ak sevice period. Additionally, the probability of custemer loss in queue models with finite capacity as spectral cases for GI / M 1/n-l queue models have been calculated by the formula found. Besides, the avarage interval between the mean first loss moment has been calculated by the determinant of a{),al,...,an_[ elements. The asimptotic analysis of the lost customer stream has beeen also done, and it has been verified that the customer stream, which disappears in a - » 0 approaches to Poisson stream when a service period is longer than the inter-arrival time. Last in that chapter it has been also proved that the probability of customer loss in the GI / M 1/0 queue model has been minimum if the inter-arrival time is kept fixed. In the chapter of practice, the findings taken from Samsun stock business administration in accordance with GI / M 1/3 model with three waiting capacity and service variant that is a spectral case of the queue model in question have been evaluated. The probability of the customer loss in Samsun stock business administration, and the avarage interval between mean first loss moment and adjacent loss moment have been calculated by the help of the formula found. Keywords: The queue model with finite capacity, the probability of customer loss, semi-Markov process
Description
Tez (doktora) -- Ondokuz Mayıs Üniversitesi, 2004
Libra Kayıt No: 23126
Libra Kayıt No: 23126
Keywords
Citation
WoS Q
Scopus Q
Source
Volume
Issue
Start Page
End Page
59
