A guide to simulation (2nd ed.)
A guide to simulation (2nd ed.)
Restart: a straightforward method for fast simulation of rare events
WSC '94 Proceedings of the 26th conference on Winter simulation
Rare event simulation in stochastic models
WSC '95 Proceedings of the 27th conference on Winter simulation
Rare event simulation of delay in packet switching networks using DPR-based splitting
Proceedings of the 31st conference on Winter simulation: Simulation---a bridge to the future - Volume 1
Implementation of Importance Splitting Techniques in Stochastic Petri Net Package
TOOLS '00 Proceedings of the 11th International Conference on Computer Performance Evaluation: Modelling Techniques and Tools
ANSS '03 Proceedings of the 36th annual symposium on Simulation
On the efficiency of RESTART for multidimensional state systems
ACM Transactions on Modeling and Computer Simulation (TOMACS)
Computer simulation of rare event for M/M/1/N queuing system
CompSysTech '07 Proceedings of the 2007 international conference on Computer systems and technologies
The Journal of Supercomputing
The rare event simulation method RESTART: efficiency analysis and guidelines for its application
Network performance engineering
ISPA'06 Proceedings of the 4th international conference on Parallel and Distributed Processing and Applications
| Hi-index | 0.01 |
Using the LRE-algorithm (LRE: Limited Relative Error) for the evaluation of simulated data yields the sta tionary distribution function of an investigated random sequence and additionally the so-called local correlation coefficient, which represents relevant correlation evidence to be included in the error measure for controlling the simulation run time. In this paper a simplified LRE-algorithm is used to evaluate discrete sequences like the occupancy of finite buffer queueing systems G/G/1/N. It is shown how this algorithm is combined with the RESTART-method for an efficient rare event simulation. A multi-stage RESTART/LRE-algorithm has been implemented as part of a stochastic simulation system and its performance has been verified by extensive simulations of the reference system M/M/1/N, whose properties including the local correlation coefficient can be described by analytical formulas. Approximate formulas for the optimal number of stages and number of trials are given. The new algorithm has been successfully applied to several finite buffer queueing systems, including the SSMP(2)/D/1/N system with a correlated input stream, in order to gain by simulation the distribution function of the occupancy including very low loss probabilities in the order of 10/sup -9/ and even 10/sup -21/, which are relevant for the performance of switching nodes in the ATM broadband network.