Method of TTTensor.
Compute the incidence matrix of a Tensor-Train-based mdss model.
Source:
src/tensor/@TTTensor/computeIncidenceMatrix.m
I = computeIncidenceMatrix(obj, sys)computeIncidenceMatrix computes the incidence matrix
describing the variable dependencies of the mdss equations represented
in Tensor-Train (TT) format.
Two computation methods are available: The default “exact” method determines the incidence matrix by enumerating all monomials represented by the Tensor Train. This approach guarantees the exact incidence matrix and is recommended for model analysis and decomposition.
Alternatively, the “jacobian” method estimates the incidence matrix by evaluating the Jacobian at multiple randomly selected operating points. Although this approach is computationally faster, it may miss variable dependencies for larger or highly coupled systems and therefore should primarily be used for quick approximations.
method |
Description |
|---|---|
"exact" (default) |
Computes the incidence matrix by exact monomial enumeration through the TT cores. Always produces the exact incidence matrix and is the recommended approach. |
"jacobian" |
Estimates the incidence matrix by evaluating the Jacobian at multiple randomly selected operating points. Computationally faster but may fail to detect all variable dependencies, particularly for larger or more complex systems. |
| Argument | Description |
|---|---|
obj |
TTTensor
object. |
sys |
mdss model
object. |
method |
"exact" (default) or "jacobian". |
tol |
Numerical threshold used to detect nonzero entries (default
1e-9). |
| Output | Description |
|---|---|
I |
Sparse logical incidence matrix. |
TTTensor · computeSPhi
· computeJacobianMatrixMdss
· mdss
MyToolbox Documentation | Generated automatically by CI/CD pipeline