pub fn compress_jacobian_pattern(
sparsity: &CsrArray<f64>,
) -> Result<CompressedJacobianPattern, OptimizeError>
Expand description
Compresses a sparse Jacobian pattern for more efficient finite differencing
Uses a simplified Curtis-Powell-Reid algorithm with graph coloring to group columns that can be computed simultaneously without interference.
§Arguments
sparsity
- Sparse matrix representing the Jacobian sparsity pattern
§Returns
- Compressed sparsity pattern and compression matrices (original, B, C) where B is the column compression matrix and C is the row compression matrix