Quadratic Programming
Contents
10.12. Quadratic Programming#
10.12.1. Quadratic Functions#
Definition 10.41 (Quadratic function)
A function 
where 
The matrix 
Remark 10.8 (Gradient and Hessian of a quadratic function)
Let
be a quadratic function. Then, the gradient is given by:
And, the Hessian is given by:
See Example 5.8 and Example 5.13 for reference.
10.12.1.1. Stationary Points#
Theorem 10.86 (Stationary points of quadratic functions)
Let a quadratic function 
where 
- If - If - If 
Proof. (1) is a direct implication of the fact that 
(2) We are given that 
- Thus, - By Theorem 8.4, if - By first part, 
(3) We are given that 
- Then, - Hence, - By parts (1) and (2), it is the unique (hence strict) global minimizer of 
(4) We know that strict global minimum point of 
10.12.1.2. Coerciveness#
Theorem 10.87 (Coerciveness of quadratic functions)
Let a quadratic function 
where 
Proof. Assume that 
- Then, all eigenvalues of - Let - Then, - Thus, 
- We can see that - Thus, 
Now, assume that 
- We need to show that - Thus, all eigenvalues of - For contradiction, assume that an eigenvalue of - Let - Then, for any - Clearly, - Thus, it contradicts the hypothesis that - We now consider the possibility where there is a 0 eigenvalue. 
- Then, there exists a normalized eigenvector - Then, for any - If - If - If - In all the three cases, - Thus, - Hence, the eigenvalues of - Hence, 
10.12.1.3. Nonnegative Quadratics#
It is useful to work with quadratic functions which are nonnegative
on the entire 
The basic quadratic form 
For the general quadratic function, we need to incorporate the
contribution from 
Theorem 10.88 (Nonnegativity of quadratic function)
Let a quadratic function 
where 
The following statements are equivalent.
Proof. Assume that (2) is true.
- Then, for every - due to positive semidefiniteness. 
- But 
- Thus, 
For the converse, assume (1) is true.
- We need to show that - We shall first show that - For contradiction, assume that - Then, there exists a negative eigenvalue - Then, for any 
- Then, - This contradicts the hypothesis that - Thus, - We now need to show that for any - This condition is equivalent to - for every - If - This is valid for every - For - By hypothesis, - Thus, - Thus, 
10.12.2. Quadratic Optimization Problems#
We consider the following possibilities:
- The objective function is a quadratic function. 
- Both the objective function as well as the inequality constraints function are quadratic. 
10.12.2.1. Quadratic Program#
Definition 10.42 (Quadratic program)
A convex optimization problem is known as a quadratic program (QP) if the objective function is a (convex) quadratic and the constraint functions are affine.
A general quadratic program has the following form:
where
10.12.2.2. Quadratically Constrained Quadratic Program#
Definition 10.43 (Quadratically constrained quadratic program)
A convex optimization problem is known as a quadratically constrained quadratic program (QCQP) if the objective function and the inequality constraint functions are (convex) quadratic while the equality constraint functions are affine.
A general quadratic program has the following form:
where