PDESpecificationBase<T>
Base Classes · AiDotNet.PhysicsInformed.PDEs
Base class for all Partial Differential Equation (PDE) specifications. Provides common functionality for PDE implementations used with Physics-Informed Neural Networks.
Implements: IPDESpecification<T>
Known derived types: AdvectionDiffusionEquation<T>, AllenCahnEquation<T>, BlackScholesEquation<T>, BurgersEquation<T>, HeatEquation<T>, KortewegDeVriesEquation<T>, LinearElasticityEquation<T>, MaxwellEquations<T>, MultiScaleReactionDiffusion<T>, NavierStokesEquation<T>, PoissonEquation<T>, SchrodingerEquation<T>, WaveEquation<T>
For Beginners
A Partial Differential Equation (PDE) describes how a quantity changes with respect to multiple variables (like space and time). This base class provides the foundation for implementing any PDE that can be solved using Physics-Informed Neural Networks.
How It Works
All PDE implementations should inherit from this class rather than implementing
IPDESpecification directly. This ensures consistent behavior and
provides access to common helper methods.
To create a new PDE, override the abstract members:
ComputeResidual- Calculate how much the PDE is violatedInputDimension- Number of independent variables (space + time)OutputDimension- Number of dependent variables (solution components)Name- Human-readable name of the PDE
Properties
| Property | Summary |
|---|---|
int DerivativeOrder { get; } |
The highest input-derivative order ComputeTapeResidual reads (1 to 3). |
int InputDimension { get; } |
|
int OutputDimension { get; } |
|
string Name { get; } |
Methods
| Method | Summary |
|---|---|
T ComputeResidual(Vector<T> inputs, Vector<T> outputs, PDEDerivatives<T> derivatives) |
|
Tensor<T> ComputeTapeResidual(Tensor<T> inputs, Tensor<T> outputs, PDEDerivatives<T> derivatives) |
Computes the PDE residual at a batch of points with tape-differentiable engine operations. |