scientific article; zbMATH DE number 1302159
approximation by spline seriesestimation of densitiesnonlinear oscillatorsnumerical integrationprobability density functionstochastic differential equations
Stochastic ordinary differential equations (aspects of stochastic analysis) (60H10) Applications of stochastic analysis (to PDEs, etc.) (60H30) Probabilistic models, generic numerical methods in probability and statistics (65C20) Random vibrations in mechanics of particles and systems (70L05) Numerical approximation of solutions of dynamical problems in solid mechanics (74H15)
This short note deals with the problem of estimating the short time joint transition probability densities \(p(y,t\mid y',t')\) of autonomous, ordinary stochastic differential equations (SDEs). Some numerical methods are briefly presented and applied to estimate these densities \(p\) of SDEs as one meets in the theory of random vibrations. The author suggests to consider the logarithm of the probability density function (PDF) to be approximated by spline series for a better replication of polynomial-type exponents of the densities. The presented technique belongs to the class of path-integral solution techniques (PIS). The author is motivated by extremely accurate numerical results for the tail behavior of joint PDF in dimensions \(d=2,3\), assuming \(\delta\)-function-type initial conditions \(p_0(y)\). Using the explicit numerical integration schemes (Euler, Milstein schemes) applied to Markovian SDEs, the author presents an explicit formula for discretized PDF of related Markov chain. Then, for the logarithm of this discretized PDF, he suggests to use cubic splines (compare also with an idea which is found in a conference paper of Naess and Moe (1997)). The overall probability density can be approximated by requiring highly accurate estimates for the local transition densities and exploiting the classical Chapman-Kolmogorov equations. The idea is useful, natural and innovative, but the presented mathematics does not give neither any proof of convergence in terms of convergence orders nor the conditions under which it could be proven. For the entire collection see [Zbl 0914.00073].
- On numerical density approximations of solutions of SDEs with unbounded coefficients
- An analytical-numerical method for fast evaluation of probability densities for transient solutions of nonlinear Itô's stochastic differential equations
- A recursive integration method for approximate solution of stochastic differential equations
- Numerical solution of Fokker-Planck equation for nonlinear stochastic dynamical systems
- An equivalent polynomial approximation technique in nonlinear structural dynamics
- Solving nonlinear dynamic stochastic models: an algorithm computing value function by simulations
- Numerical meshfree path integration method for non-linear dynamic systems
- THE NUMERICAL SOLUTION OF NONLINEAR STOCHASTIC DYNAMICAL SYSTEMS: A BRIEF INTRODUCTION
- A path integration algorithm for stochastic structural dynamic systems
- On numerical density approximations of solutions of SDEs with unbounded coefficients
This page was built for publication:
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4247425)