probnum.diffeq

Differential Equations.

This package defines common dynamical models and probabilistic solvers for differential equations.

Functions

probsolve_ivp(f, t0, tmax, y0[, df, method, …])	Solve initial value problem with Gaussian filtering and smoothing.
propose_firststep(ivp)	Propose a suitable first step that can be taken by an ODE solver.
initialize_odefilter_with_rk(f, y0, t0, …)	Initialize an ODE filter by fitting the prior process to a few steps of an approximate ODE solution computed with Scipy’s RK.
initialize_odefilter_with_taylormode(f, y0, …)	Initialize an ODE filter with Taylor-mode automatic differentiation.
perturb_lognormal(rng, step, solver_order, …)	Perturb the step with lognormally distributed noise scaled by noise-scale.
perturb_uniform(rng, step, solver_order, …)	Perturb the step with uniformly distributed noise scaled by noise-scale.

Classes

ODESolver(ivp, order)	Interface for ODESolver.
GaussianIVPFilter(ivp, prior_process, …[, …])	ODE solver that uses a Gaussian filter.
StepRule(firststep)	(Adaptive) step size rules for ODE solvers.
ConstantSteps(stepsize)	Constant step size rule for ODE solvers.
AdaptiveSteps(firststep, atol, rtol[, …])	Adaptive step size selection using proportional control.
ODESolution(locations, states[, derivatives])	ODE solution.
KalmanODESolution(kalman_posterior)	Gaussian IVP filtering solution of an ODE problem.
PerturbedStepSolver(rng, solver, …)	ODE-Solver random perturbatino of the step-sizes.
PerturbedStepSolution(scales, locations, …)	Solution to the PerturbedStepSolver.
WrappedScipyRungeKutta(solver)	Wrapper for Runge-Kutta methods from Scipy, implements the stepfunction and dense output.

Class Inheritance Diagram