PoissonOperatorBenchmarkOptions
Options & Configuration · AiDotNet.PhysicsInformed.Benchmarks
Configures the 2D Poisson operator-learning dataset that OperatorBenchmarkSuite generates: each
sample pairs a random forcing field f(x, y) with the u solving -Δu = f on the unit square.
For Beginners
The input image is a heat source spread over a plate; the output is the steady temperature it produces. A model that learns this mapping can stand in for the solver.
How It Works
The forcing is a random sine series, and u is solved by iterating to Tolerance with zero boundary
values. A neural operator learns the map f ↦ u, the standard first test for operator learning.
Constructors
| Constructor | Summary |
|---|---|
PoissonOperatorBenchmarkOptions() |
Properties
| Property | Summary |
|---|---|
double Tolerance { get; set; } |
Gets or sets the largest change of any grid value in a sweep at which the solver stops. |
int GridSize { get; set; } |
Gets or sets the number of grid points per side of the unit square. |
int MaxFrequency { get; set; } |
Gets or sets the highest sine mode per axis in the random forcing series. |
int MaxIterations { get; set; } |
Gets or sets the most solver sweeps run for each sample. |
int SampleCount { get; set; } |
Gets or sets the number of (forcing, solution) pairs generated. |
int Seed { get; set; } |
Gets or sets the seed of the random fields, so a dataset can be regenerated exactly. |
Methods
| Method | Summary |
|---|---|
void Validate() |
Throws ArgumentOutOfRangeException for any value outside its documented range. |