Optimality condition for a convex, linearly constrained problem

Theorem: optimality condition for convex, linearly constrained problems
Consider the linearly constrained optimization problem

    \[\min _x f_0(x): A x=b,\]

where f_0: \mathbf{R}^n \rightarrow \mathbf{R} is convex and differentiable, and A \in \mathbf{R}^{m \times n}, b \in \mathbf{R}^m define the equality constraints.
A point x is optimal for the above problem if and only if it satisfies

    \[A x=b, \quad \nabla f_0(x)=A^T \lambda .\]

for some vector \lambda \in \mathbf{R}^m.
Proof: Let us re-formulate the optimality condition is

    \[x \in \mathbf{X} \text { and } \forall y \in \mathbf{X}: \nabla f_0(x)^T(y-x) \geq 0 .\]

where the feasible set \mathbf{X} is now the affine set

    \[\{x: A x=b\} \text {. }\]

We can write the above as: A x=b and

    \[\forall z, \quad A z=0: \nabla f_0(x)^T z \geq 0 .\]

Since we can always flip the sign of vectors z with A z=0, we can express the above as

    \[\forall z, \quad A z=0: \nabla f_0(x)^T z=0 .\]

This means that \nabla f_0(x) should be orthogonal to the nullspace of A.
From the Fundamental theorem of linear algebra, this in turn says that \nabla f_0(x) should belong to the range of the transpose matrix, \mathbf{R}\left(A^T\right). In other words, the above condition is equivalent to

    \[\exists \lambda \in \mathbf{R}^m: \nabla f_0(x)=A^T \lambda .\]

We conclude that the optimality conditions for x to be optimal are that there exists a vector \lambda such that

    \[A x=b, \quad \nabla f_0(x)=A^T \lambda .\]

This ends our proof.

License

Hyper-Textbook: Optimization Models and Applications Copyright © by L. El Ghaoui. All Rights Reserved.

Share This Book