Μετάβαση στο περιεχόμενο. | Μετάβαση στην πλοήγηση

Σεμινάριο Επιχειρησιακής Έρευνας, Janos Sztrik

Πότε 22/05/2018
από 10:00 έως 11:00
Που Α32
Προσθήκη γεγονότος στο ημερολόγιο vCal
iCal
Σεμινάριο Τομέα Στατιστικής και Επιχειρησιακής Έρευνας.

Ημερομηνία: Τρίτη 22 Μαΐου 2018, ώρα 10:00-11:00. Αίθουσα: A32.

Τίτλος: Recent results on finite source retrial queues with collisions.

Ομιλήτης: Janos Sztrik, Professor, University of Debrecen, Debrecen, Hungary
(σε συνεργασία με τον A. Nazarov, Professor, National Research Tomsk State
University, Tomsk, Russia).

Περίληψη: Retrial queues have been widely used to model many problems
arising in telephone switching systems, call centers, info-communication
networks just to mention some. There is no queueing in these systems, that
is if an arriving customer finds the server busy it goes into the orbit
from where he makes random retrials to get service. Again, if at the
moment of arrival the server is busy he returns to the orbit and the whole
procedure repeats until the server is idle. If at the moment of arrival
from the source the server is idle the service starts immediately. It is
assumed that the customers arrive from a finite number of sources N. The
novelty of the investigations is that if the arriving customer finds the
server busy he involves into collision with the customer under service and
thus both enter into the orbit. Clearly, in this case the utilization of
the server and the throughput of the system decrease, while  waiting time,
response time, number of retrials and the number of customers in the
systems increase. The structure of the mathematical model describing the
system behavior depends on the distribution of the source times, service
times and retrial times which are supposed to be independent random
variables. The aim of the presentation is to give an overview on the
recent research carried out by the help of analytic, numeric, stochastic
simulation, asymptotic methods and to summarize the most important results
obtained on this topic. Special attention is devoted to the distribution
of number of customers at the service facility, distribution of number of
retrials and finally to the most complicated problem, namely to the
distribution of waiting and response time. Several figures will illustrate
the problems in question.
Αρχειοθέτηση