On periodic parameter identification in stochastic differential equations

• DOI: 10.3934/ipi.2019025
• Publication Date: 2019-03-25T04:29:35Z
• Authors: Pingping Niu; ; Shuai Lu; Jin Cheng
• Keywords: Regularization

Periodic parameters are common and important in stochastic differential equations (SDEs) arising many contemporary scientific engineering fields involving dynamical processes. These include the damping coefficient, volatility or diffusion coefficient possibly an external force. Identification of these periodic allows a better understanding processes their hidden intermittent instability. Conventional approaches usually assume that one is known focus on recovery rest parameters. By introducing decorrelation time calculating standard Gaussian statistics (mean, variance) explicitly for scalar Langevin with parameters, we propose parameter identification approach to simultaneously recovering all by observing single trajectory SDEs. Such summarized form regularization schemes noisy operators right-hand sides further extended SDEs which indirectly observed other random Numerical examples show our performs well stable weakly unstable regimes but may fail strongly regime induced strong instability itself.