In this paper, we generalize Freidlin and Wentzell machinery [7] to the case of Glauber dynamics, which is still a stochastic dynamics even as the temperature vanishes. We first consider the probability to exit from a...In this paper, we generalize Freidlin and Wentzell machinery [7] to the case of Glauber dynamics, which is still a stochastic dynamics even as the temperature vanishes. We first consider the probability to exit from an attractive basin and enter into another one in the limit as the temperature goes to zero. The first exiting time to escape from an attractive basin of an attractor in which the process starts is estimated. Then we show conclusions above in the case of exiting from a region formed by geveral attractive basins. The unpredictability of the exiting time is proved.展开更多
Multiple stability for two-dimensional delayed recurrent neural networks with piecewise linear activation flmctions of 2r （r 〉 1） corner points is studied. Sufficient conditions are established for checking the exi...Multiple stability for two-dimensional delayed recurrent neural networks with piecewise linear activation flmctions of 2r （r 〉 1） corner points is studied. Sufficient conditions are established for checking the existence of （2r ＋ 1）2 equilibria in delayed recurrent neural networks. Under these conditions, （r ＋ 1）2 equilibria are locally exponentially stable, and （2r＋ 1）2 -（r ＋ 1）2 -r2 equilibria are unstable. Attractive basins of stable equilibria are estimated, which are larger than invariant sets derived by decomposing state space. One example is provided to illustrate the effectiveness of our results.展开更多
A family of irreducible Markov chains on a finite state space is considered as an exponential perturbation of a reducible Markov chain. This is a generalization of the Freidlin-Wentzell theory, motivated by studies of...A family of irreducible Markov chains on a finite state space is considered as an exponential perturbation of a reducible Markov chain. This is a generalization of the Freidlin-Wentzell theory, motivated by studies of stochastic Ising models, neural network and simulated annealing. It is shown that the metastability is a universal feature for this wide class of Markov chains. The metastable states are simply those recurrent states of the reducible Markov chain. Higher level attractors, related attractive basins and their pyramidal structure are analysed. The logarithmic asymptotics of the hitting time of various sets are estimated. The hitting time over its mean converges in law to the unit exponential distribution.展开更多
文摘In this paper, we generalize Freidlin and Wentzell machinery [7] to the case of Glauber dynamics, which is still a stochastic dynamics even as the temperature vanishes. We first consider the probability to exit from an attractive basin and enter into another one in the limit as the temperature goes to zero. The first exiting time to escape from an attractive basin of an attractor in which the process starts is estimated. Then we show conclusions above in the case of exiting from a region formed by geveral attractive basins. The unpredictability of the exiting time is proved.
基金Project supported by the National Natural Science Foundation of China (Grant Nos. 50977008, 61034005, and 61074073)the National Basic Research Program of China (Grant No. 2009CB320601)+1 种基金the Program for New Century Excellent Talents in Universities of China (Grant No. NCET-10-0306)the Fundamental Research Funds for the Central Universities of China(Grant Nos. N110604005 and N110504001)
文摘Multiple stability for two-dimensional delayed recurrent neural networks with piecewise linear activation flmctions of 2r （r 〉 1） corner points is studied. Sufficient conditions are established for checking the existence of （2r ＋ 1）2 equilibria in delayed recurrent neural networks. Under these conditions, （r ＋ 1）2 equilibria are locally exponentially stable, and （2r＋ 1）2 -（r ＋ 1）2 -r2 equilibria are unstable. Attractive basins of stable equilibria are estimated, which are larger than invariant sets derived by decomposing state space. One example is provided to illustrate the effectiveness of our results.
基金Project partially supported by the National Natural Science Foundation of China and a Postdoctoral Fellowship of the State Education Commission.
文摘A family of irreducible Markov chains on a finite state space is considered as an exponential perturbation of a reducible Markov chain. This is a generalization of the Freidlin-Wentzell theory, motivated by studies of stochastic Ising models, neural network and simulated annealing. It is shown that the metastability is a universal feature for this wide class of Markov chains. The metastable states are simply those recurrent states of the reducible Markov chain. Higher level attractors, related attractive basins and their pyramidal structure are analysed. The logarithmic asymptotics of the hitting time of various sets are estimated. The hitting time over its mean converges in law to the unit exponential distribution.
