We generalize the decomposition method of the finite Markov chains for Poincare inequality in Jerrum et al.（Ann.Appl.Probab.,14,1741-1765（2004）） to the reversible continuous-time Markov chains.And inductively,we g...We generalize the decomposition method of the finite Markov chains for Poincare inequality in Jerrum et al.（Ann.Appl.Probab.,14,1741-1765（2004）） to the reversible continuous-time Markov chains.And inductively,we give the lower bound of spectral gap for the ergodic open Jackson network by the decomposition method and the symmetrization procedure.The upper bound of the spectral gap is also presented.展开更多
The asymptotic variability analysis is studied for multi-server generalized Jackson network. It is characterized by law of the iterated logarithm （LIL）, which quantifies the magnitude of asymptotic stochastic fluctu...The asymptotic variability analysis is studied for multi-server generalized Jackson network. It is characterized by law of the iterated logarithm （LIL）, which quantifies the magnitude of asymptotic stochastic fluctuations of the stochastic processes compensated by their deterministic fluid limits. In the overloaded （OL） case, the asymptotic variability is studied for five performance measures： queue length, workload, busy time, idle time and number of departures. The proof is based on strong approximations, which approximate discrete performance processes with （reflected） Brownian motions. We conduct numerical examples to provide insights on these LIL results.展开更多
A general Jackson network （GJN） with infinite supply of work is considered. By fluid limit model, the author finds that the Markov process describing the dynamics of the GJN with infinite supply of work is positive ...A general Jackson network （GJN） with infinite supply of work is considered. By fluid limit model, the author finds that the Markov process describing the dynamics of the GJN with infinite supply of work is positive Harris recurrent if the corresponding fluid model is stable. Furthermore, the author proves that the fluid model is stable if the usual traffic condition holds.展开更多
Using a bounding technique, we prove that the fluid model of generalized Jackson network （GJN） with vacations is the same as a GJN without vacations, which means that vacation mechanism does not affect the dynamic p...Using a bounding technique, we prove that the fluid model of generalized Jackson network （GJN） with vacations is the same as a GJN without vacations, which means that vacation mechanism does not affect the dynamic performance of GJN under fluid approximation. Furthermore, in order to present the impact of vacation on the performance of GJN, we show that exponential rate of convergence for fluid approximation only holds for large N, which is different from a GJN without vacations. The results on fluid approximation and convergence fate are embodied by the queue length, workload, and busy time processes.展开更多
The relationship between the order of approximation by neural network based on scattered threshold value nodes and the neurons involved in a single hidden layer is investigated. The results obtained show that the degr...The relationship between the order of approximation by neural network based on scattered threshold value nodes and the neurons involved in a single hidden layer is investigated. The results obtained show that the degree of approximation by the periodic neural network with one hidden layer and scattered threshold value nodes is increased with the increase of the number of neurons hid in hidden layer and the smoothness of excitation function.展开更多
基金Supported in part by 985 Project973 Project(Grant No.2011CB808000)+2 种基金NSFC(Grant No.11131003)SRFDP(Grant No.20100003110005)the Fundamental Research Funds for the Central Universities
文摘We generalize the decomposition method of the finite Markov chains for Poincare inequality in Jerrum et al.（Ann.Appl.Probab.,14,1741-1765（2004）） to the reversible continuous-time Markov chains.And inductively,we give the lower bound of spectral gap for the ergodic open Jackson network by the decomposition method and the symmetrization procedure.The upper bound of the spectral gap is also presented.
基金Supported by the National Natural Science Foundation of China(No.11471053)
文摘The asymptotic variability analysis is studied for multi-server generalized Jackson network. It is characterized by law of the iterated logarithm （LIL）, which quantifies the magnitude of asymptotic stochastic fluctuations of the stochastic processes compensated by their deterministic fluid limits. In the overloaded （OL） case, the asymptotic variability is studied for five performance measures： queue length, workload, busy time, idle time and number of departures. The proof is based on strong approximations, which approximate discrete performance processes with （reflected） Brownian motions. We conduct numerical examples to provide insights on these LIL results.
文摘A general Jackson network （GJN） with infinite supply of work is considered. By fluid limit model, the author finds that the Markov process describing the dynamics of the GJN with infinite supply of work is positive Harris recurrent if the corresponding fluid model is stable. Furthermore, the author proves that the fluid model is stable if the usual traffic condition holds.
文摘Using a bounding technique, we prove that the fluid model of generalized Jackson network （GJN） with vacations is the same as a GJN without vacations, which means that vacation mechanism does not affect the dynamic performance of GJN under fluid approximation. Furthermore, in order to present the impact of vacation on the performance of GJN, we show that exponential rate of convergence for fluid approximation only holds for large N, which is different from a GJN without vacations. The results on fluid approximation and convergence fate are embodied by the queue length, workload, and busy time processes.
文摘The relationship between the order of approximation by neural network based on scattered threshold value nodes and the neurons involved in a single hidden layer is investigated. The results obtained show that the degree of approximation by the periodic neural network with one hidden layer and scattered threshold value nodes is increased with the increase of the number of neurons hid in hidden layer and the smoothness of excitation function.
