Skip to main content

5.7 Quadratic Programming

 

Consider the case of a convex quadratic objective function, where the constraints are affine i.e.,

$min_{x \in \mathbb{R}^d} \quad \frac{1}{2}x^TQx+c^Tx$

subject to $Ax \le b$

Where $A \in \mathbb{R}^{m \times d},b \in \mathbb{R}^m$ and $c \in \mathbb{R}^d$. The square symmetric matrix $ Q \in \mathbb{R}^{d \times d}$ is positive definite, and therefore the objective function is convex.This is known as a quadratic program. Observe that it has $d$ variables and $m$ linear constraints.

Consider the quadratic program of two variables

$min_{x \in \mathbb{R}^2}\frac{1}{2}\begin{bmatrix}
x_1\\
x_2
\end{bmatrix}^T\begin{bmatrix}
2 & 1\\
1 & 4
\end{bmatrix}
\begin{bmatrix}
x_1\\
x_2
\end{bmatrix}+\begin{bmatrix}
5\\
3
\end{bmatrix}^T\begin{bmatrix}
x_1\\
x_2
\end{bmatrix}$

subject to
$\begin{bmatrix}
1 & 0\\
-1& 0\\
0& 1\\
0& -1
\end{bmatrix}\begin{bmatrix}
x_1\\
x_2
\end{bmatrix} \le \begin{bmatrix}
1\\
1\\
1\\
1
\end{bmatrix}$
This is illustrated in figure

The objective function is quadratic with a positive semidefinite matrix $Q$, resulting in elliptical contour lines. The optimal value must lie in the shaded (feasible) region, and is indicated by the star.

The Lagrangian is given by

$L(x,\lambda)=\frac{1}{2}x^TQx+c^Tx+\lambda^T(Ax-b)$
$\quad=\frac{1}{2}x^TQx+(c+A^T\lambda)^Tx-\lambda^Tb$

where again, we have re arranged the terms.Taking the derivative of $L(x,\lambda)$ with respect to $x$ and setting it to zero gives

$Qx+(c+A^T\lambda)=0$

Assuming that $Q$ is invertible, we get
$x = -Q^{-1}(c + A^T\lambda)$

Substituting  into the primal Lagrangian $L(x,\lambda)$, we get the dual Lagrangian
$D(\lambda)=-\frac{1}{2}(c+A^T\lambda)^TQ^{-1}(c+A^T\lambda)-\lambda^Tb$

Therefore the dual optimization problem is given by

$max_{\lambda \in \mathbb{R}^m}\quad  -\frac{1}{2}(c+A^T\lambda)^TQ^{-1}(c+A^T\lambda)-\lambda^Tb$

subject to $\lambda \ge 0$

Example: Consider the above problem

Provide necessary and sufficient conditions under which a quadratic optimization problem be written as a linear least squares problem. 

Solution 
Consider $min_{x \in R^n} \frac{1}{2} x^ TQx + g^T x$, then the necessary and sufficient condition is $Q$ is PSD and $g \in Ran(Q)$. Indeed, if $Q$ is PSD, then $Q$ has a Cholesky factorization $Q = LL^T$ where $L \in R^{ n×k}$ with $k = rank(Q)$. Since $g \in Ran(Q) = Ran(L)$, there is a vector $b \in R^ k$ such that $−g = Lb$. 
Then $\frac{1}{2} x^TQx + g ^T x = \frac{1}{2} x^TLL^T x − (Lb)^ T x$
  $ = || \frac{1}{2} (L^T x) ^T (L^T x) − b^T (L^T x) + \frac{1}{2} b^T b || − \frac{1}{2} b^T b $
$= \frac{1}{2} ||L^T x − b||^T_ 2 −\frac{1}{2} b^Tb$

Comments

Popular posts from this blog

Mathematics for Machine Learning- CST 284 - KTU Minor Notes - Dr Binu V P

  Introduction About Me Syllabus Course Outcomes and Model Question Paper University Question Papers and Evaluation Scheme -Mathematics for Machine learning CST 284 KTU Overview of Machine Learning What is Machine Learning (video) Learn the Seven Steps in Machine Learning (video) Linear Algebra in Machine Learning Module I- Linear Algebra 1.Geometry of Linear Equations (video-Gilbert Strang) 2.Elimination with Matrices (video-Gilbert Strang) 3.Solving System of equations using Gauss Elimination Method 4.Row Echelon form and Reduced Row Echelon Form -Python Code 5.Solving system of equations Python code 6. Practice problems Gauss Elimination ( contact) 7.Finding Inverse using Gauss Jordan Elimination  (video) 8.Finding Inverse using Gauss Jordan Elimination-Python code Vectors in Machine Learning- Basics 9.Vector spaces and sub spaces 10.Linear Independence 11.Linear Independence, Basis and Dimension (video) 12.Generating set basis and span 13.Rank of a Matrix 14.Linear Mapping...

4.3 Sum Rule, Product Rule, and Bayes’ Theorem

 We think of probability theory as an extension to logical reasoning Probabilistic modeling  provides a principled foundation for designing machine learning methods. Once we have defined probability distributions corresponding to the uncertainties of the data and our problem, it turns out that there are only two fundamental rules, the sum rule and the product rule. Let $p(x,y)$ is the joint distribution of the two random variables $x, y$. The distributions $p(x)$ and $p(y)$ are the corresponding marginal distributions, and $p(y |x)$ is the conditional distribution of $y$ given $x$. Sum Rule The addition rule states the probability of two events is the sum of the probability that either will happen minus the probability that both will happen. The addition rule is: $P(A∪B)=P(A)+P(B)−P(A∩B)$ Suppose $A$ and $B$ are disjoint, their intersection is empty. Then the probability of their intersection is zero. In symbols:  $P(A∩B)=0$  The addition law then simplifies to: $P(...

5.1 Optimization using Gradient Descent

Since machine learning algorithms are implemented on a computer, the mathematical formulations are expressed as numerical optimization methods.Training a machine learning model often boils down to finding a good set of parameters. The notion of “good” is determined by the objective function or the probabilistic model. Given an objective function, finding the best value is done using optimization algorithms. There are two main branches of continuous optimization constrained and unconstrained. By convention, most objective functions in machine learning are intended to be minimized, that is, the best value is the minimum value. Intuitively finding the best value is like finding the valleys of the objective function, and the gradients point us uphill. The idea is to move downhill (opposite to the gradient) and hope to find the deepest point. For unconstrained optimization, this is the only concept we need,but there are several design choices. For constrained optimization, we need to intr...