Name
Abstract | ||
|---|---|---|
Queuing theory strives to provide analytical solutions to a number of queuing prob- lems. Unfortunately, closed analytical expressions can not be derived for every type of queuing system. Discrete event simulation on the other hand can find solutions to any kind of queuing problem, but the results are stochastic and only specific to a single parameterization. Proxel simulation can bridge the gap between these two methods by providing accurate deterministic results of the performance measures of a wide range of queuing systems. Furthermore one can solve some stiff problems far more quickly than when using discrete event simulation. The paper presents the general approach and a tool implementing the Proxel-based simulation of queuing systems. Experiments validate the method and show the range of applicability. 1 Motivation Queuing Theory is one of the oldest branches of simulation. Still today many communi- cations problems are analyzed and solved using basic queuing theory mechanisms. The basic components of a queuing system are queues, servers and customers. The goal of the analysis of such a queuing system is finding analytical expressions for such performance measures as steady state queue length, throughput and utilization. The more complex the models get, the more difficult it is to derive analytical expressions for the performance measures. For a number of queuing problems it is not even possible to get these analytical expressions. Most of these difficult problems involve general distributions for the arrival and service distributions or multiple server environments. It would be desirable to have a tool that yields deterministic results for these models' perfor- mance, if no analytical expression is available. Discrete event simulation is an alternative to queuing analysis. Unfortunately, it has the drawback of being stochastic and only providing results for a specific parameterized problem, compared to the general solution of queuing theory. Therefore the results are not comparable and often not of use to queuing analysts. Using Proxels, the simulation of queuing models becomes more attractive and useful to queuing analysts. Proxels still can only provide results for a specific queuing system, but the results are deterministic and can be obtained to an arbitrary accuracy. One can obtain quick and dirty results for new queuing problems, that have not been analyzed thoroughly yet, or increase the quality of results for problems that can not be solved analytically by common queuing theory. |
Year | Venue | Keywords |
|---|---|---|
2007 | SimVis | analytic solution,queuing theory,queuing system,discrete event simulation,steady state |
Field | DocType | Citations |
Mathematical optimization,Parametrization,Expression (mathematics),Simulation,Computer science,Queueing theory,Cumulative flow diagram,Queue management system,Discrete event simulation,Fair queuing | Conference | 1 |
PageRank | References | Authors |
0.40 | 1 | 3 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Claudia Krull | 1 | 9 | 4.37 |
| graham horton | 2 | 1 | 0.40 |
| ottovonguericke universitat magdeburg | 3 | 3 | 0.89 |