SciML / BoundaryValueDiffEq.jl

Boundary value problem (BVP) solvers for scientific machine learning (SciML)
Other
44 stars 36 forks source link
bvp differential-equations differentialequations gpu neural-bvp neural-differential-equations neural-ode scientific-machine-learning sciml

BoundaryValueDiffEq

Join the chat at https://julialang.zulipchat.com #sciml-bridged Global Docs

Build Status codecov Package Downloads Aqua QA

ColPrac: Contributor's Guide on Collaborative Practices for Community Packages SciML Code Style

BoundaryValueDiffEq.jl is a component package in the DifferentialEquations ecosystem. It holds the boundary value problem solvers and utilities. While completely independent and usable on its own, users interested in using this functionality should check out DifferentialEquations.jl.

API

BoundaryValueDiffEq.jl is part of the SciML common interface, but can be used independently of DifferentialEquations.jl. The only requirement is that the user passes a BoundaryValueDiffEq.jl algorithm to solve. For example, we can solve the BVP tutorial from the documentation using the MIRK4() algorithm:

using BoundaryValueDiffEq
tspan = (0.0, pi / 2)
function simplependulum!(du, u, p, t)
    θ = u[1]
    dθ = u[2]
    du[1] = dθ
    du[2] = -9.81 * sin(θ)
end
function bc!(residual, u, p, t)
    residual[1] = u[:, end ÷ 2][1] + pi / 2
    residual[2] = u[:, end][1] - pi / 2
end
prob = BVProblem(simplependulum!, bc!, [pi / 2, pi / 2], tspan)
sol = solve(prob, MIRK4(), dt = 0.05)

Available Solvers

For the list of available solvers, please refer to the DifferentialEquations.jl BVP Solvers page. For options for the solve command, see the common solver options page.

Controlling Precompilation

Precompilation can be controlled via Preferences.jl

To set these preferences before loading the package, do the following (replacing PrecompileShooting with the preference you want to set, or pass in multiple pairs to set them together):

using Preferences, UUIDs
Preferences.set_preferences!(
    UUID("764a87c0-6b3e-53db-9096-fe964310641d"), "PrecompileShooting" => false)

Running Benchmarks Locally

We include a small set of benchmarks in the benchmarks folder. These are not extensive and mainly used to track regressions during development. For more extensive benchmarks, see the SciMLBenchmarks repository.

To run benchmarks locally install AirspeedVelocity.jl and run the following command in the package directory:

benchpkg BoundaryValueDiffEq --rev="master,<git sha for your commit>" --bench-on="master"