Derived Matrices¶

Residual Jacobians (dRdX, dRdU)¶

class kona.linalg.matrices.common.dRdX(primal=None, state=None, transposed=False)[source]¶

Partial jacobian of the system residual with respect to primal variables.

class kona.linalg.matrices.common.dRdU(primal=None, state=None, transposed=False)[source]¶

Partial jacobian of the system residual with respect to state variables.

solve(rhs_vec, solution, rel_tol=1e-08)[source]¶

Performs a linear solution with the provided right hand side.

If the transposed matrix object is used, and the right hand side vector is None, then this function performs an adjoint solution.

Parameters:	rhs_vec (StateVector or None) – Right hand side vector for solution. rel_tol (float) – Solution tolerance. solution (StateVector) – Vector where the result should be stored.
Returns:	Convergence flag.
Return type:	bool

class kona.linalg.matrices.common.dCdX(primal=None, state=None, transposed=False)[source]¶

Combined partial constraint jacobian matrix that can do both equality and inequality products depending on what input vectors are provided.

class kona.linalg.matrices.common.dCdU(primal=None, state=None, transposed=False)[source]¶

Combined partial constraint jacobian matrix that can do both equality and inequality products depending on what input vectors are provided.

class kona.linalg.matrices.common.dCEQdX(primal=None, state=None, transposed=False)[source]¶

Partial jacobian of the equality constraints with respect to design vars.

class kona.linalg.matrices.common.dCEQdU(primal=None, state=None, transposed=False)[source]¶

Partial jacobian of the equality constraints with respect to state vars.

class kona.linalg.matrices.common.dCINdX(primal=None, state=None, transposed=False)[source]¶

Partial jacobian of the inequality constraints with respect to design vars.

class kona.linalg.matrices.common.dCINdU(primal=None, state=None, transposed=False)[source]¶

Partial jacobian of the inequality constraints with respect to state vars.