Multicanonical Monte Carlo (MMC) is a simulation-acceleration technique for the estimation of the statistical distribution of a desired system output variable, given the known distribution of the system input variables. MMC, similarly to the powerful and well-studied method of importance sampling (IS) [1], is a useful method to efficiently simulate events occurring with probabilities smaller than ~ 10^ − 6, such as bit error rate (BER) and system outage probability. Modern telecommunications systems often employ forward error correcting (FEC) codes that allow pre-decoded channel error rates higher than 10 ^− 3; these systems are well served by traditional Monte-Carlo error counting. MMC and IS are, nonetheless, fundamental tools to both understand the statistics of the decision variable (as well as of any physical parameter of interest) and to validate any analytical or semianalytical BER calculation model. Several examples of such use will be provided in this chapter. As a case in point, outage probabilities are routinely below 10^ − 6, a sweet spot where MMC and IS provide the most efficient (sometimes the only) solution to estimate outages.
Multicanonical Monte Carlo for Simulation of Optical Links / Bononi, Alberto; L. A., Rusch. - (2011), pp. 373-414. [10.1007/978-1-4419-8139-4_10]
Multicanonical Monte Carlo for Simulation of Optical Links
BONONI, Alberto
Formal Analysis
;
2011-01-01
Abstract
Multicanonical Monte Carlo (MMC) is a simulation-acceleration technique for the estimation of the statistical distribution of a desired system output variable, given the known distribution of the system input variables. MMC, similarly to the powerful and well-studied method of importance sampling (IS) [1], is a useful method to efficiently simulate events occurring with probabilities smaller than ~ 10^ − 6, such as bit error rate (BER) and system outage probability. Modern telecommunications systems often employ forward error correcting (FEC) codes that allow pre-decoded channel error rates higher than 10 ^− 3; these systems are well served by traditional Monte-Carlo error counting. MMC and IS are, nonetheless, fundamental tools to both understand the statistics of the decision variable (as well as of any physical parameter of interest) and to validate any analytical or semianalytical BER calculation model. Several examples of such use will be provided in this chapter. As a case in point, outage probabilities are routinely below 10^ − 6, a sweet spot where MMC and IS provide the most efficient (sometimes the only) solution to estimate outages.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
