View source for Quadratic constrained quadratic programming

Author: Jialiang Wang (jw2697), Jiaxin Zhang (jz2289), Wenke Du (wd275), Yufan Zhu (yz2899), David Berroa (deb336) (ChemE 6800 Fall 2024)

Stewards: Nathan Preuss, Wei-Han Chen, Tianqi Xiao, Guoqing Hu

==Introduction==
'''Quadratic programming (QP)''' is one of the oldest topics in the field of optimization that researchers have studied in the twentieth century. The basic QP, where the objective function is quadratic and constraints are linear, paved the way for other forms, such as QCQPs, which also have quadratic constraints. QCQPs emerged as [[wikipedia:Mathematical_optimization|optimisation theory]] grew to address more realistic, [https://zhengy09.github.io/ECE285/lectures/L6.pdf complex problems] of non-linear objectives and constraints.

A '''[[wikipedia:Quadratically_constrained_quadratic_program|Quadratically Constrained Quadratic Program (QCQP)]]''' can be defined as an optimization problem where the objective function and the constraints are quadratic. In particular, the issue involves optimizing (or minimizing) a convex quadratic function of decision variables with quadratic constraints. This class of problems is well suited to finance (Zenios,1993), engineering, machine learning, and agriculture because it is easier to model the relationship between variables using quadratic functions.

* The desire to study QCQPs stems from the fact that they can be used to model practical optimization problems that involve stochasticity in risk, resources, production, and decision-making. For example, in agriculture, using QCQPs can be useful in determining the best crop to grow based on the expected profits and the uncertainties of price changes and unfavourable weather conditions (Floudas,1995).
* In finance, the QCQPs are applied in the portfolio's construction to maximize the portfolio's expected returns and the covariance between the assets. It is crucial to comprehend QCQPs and the ways to solve them, such as '''[[wikipedia:Karush–Kuhn–Tucker_conditions|KKT (Karush-Kuhn Tucker)]]''' conditions and '''[[wikipedia:Software-defined_perimeter|SDP (Semidefinite Programming)]]''' relaxations, to solve problems that linear models cannot effectively solve (Bao & Sahinidis,2011).
{| class="wikitable"
|+
!Name
!'''Brief info'''
|-
|'''KKT(Karush-Kuhn-Tucker)'''
|KKT is a mathematical optimization method used to solve constrained optimization problems(Ghojogh, Karray et al,2021). 
It builds upon the method of Lagrange multipliers by introducing necessary conditions for optimality that incorporate primal and dual variables. 
The KKT conditions include stationarity, primal feasibility, dual feasibility, and complementary slackness, making it particularly effective for solving problems with nonlinear constraints.
|-
|'''SDP (Semidefinite Programming)'''
|SDP reformulates a QCQP problem into a semidefinite programming relaxation (Freund,2004). 
By “lifting” the problem to a higher-dimensional space and applying SDP relaxation, this approach provides a tractable way to solve or approximate solutions to non-convex QCQP problems. 
It is widely used in areas where global optimization or approximations to non-convex issues are necessary.
|}
In general, analyzing QCQPs is important in order to apply knowledge-based decision-making and enhance the performance and stability of optimization methods in different fields (Zenios,1993).

==Algorithm Discussion==



Selected Solvers:
{| class="wikitable"
|+
!Name
!Brief Info
|-
|
|
|-
|
|
|-
|
|
|}

== Numerical Example 1 (KKT Approach) ==

Consider the following '''Quadratically Constrained Quadratic Programming (QCQP)''' problem to gain a better understanding:

<math>
\begin{aligned}
\text{minimize} \quad & f_0(x) = (x_1 - 2)^2 + x_2^2 \\
\text{subject to} \quad & f_1(x) = x_1^2 + x_2^2 - 1 \leq 0, \\
& f_2(x) = (x_1 - 1)^2 + x_2^2 - 1 \leq 0.
\end{aligned}
</math>

=== Step 1: Formulate the Lagrangian ===

The Lagrangian <math>L</math> combines the objective function and the constraints, each multiplied by a Lagrange multiplier <math>\lambda_i</math>:

<math>
L(x, \lambda_1, \lambda_2) = (x_1 - 2)^2 + x_2^2 + \lambda_1 (x_1^2 + x_2^2 - 1) + \lambda_2 \left( (x_1 - 1)^2 + x_2^2 - 1 \right).
</math>

For each constraint:

- '''Complementary Slackness''':
  <math>
  \lambda_i \geq 0, \quad \lambda_i f_i(x) = 0, \quad \text{for } i = 1, 2.
  </math>

- '''Primal Feasibility''':
  <math>
  f_i(x) \leq 0 \quad \text{for } i = 1, 2.
  </math>

=== Step 2: Compute the Gradient of the Lagrangian ===

Compute the partial derivatives with respect to <math>x_1</math> and <math>x_2</math>:

- '''Partial Derivative with respect to <math>x_1</math>''':
  <math>
  \frac{\partial L}{\partial x_1} = 2(x_1 - 2) + 2\lambda_1 x_1 + 2\lambda_2 (x_1 - 1).
  </math>

- '''Partial Derivative with respect to <math>x_2</math>''':
  <math>
  \frac{\partial L}{\partial x_2} = 2x_2 + 2\lambda_1 x_2 + 2\lambda_2 x_2.
  </math>

=== Step 3: Stationarity Conditions ===

Set the gradients to zero:

- '''Equation (1)''':
  <math>
  2(x_1 - 2) + 2\lambda_1 x_1 + 2\lambda_2 (x_1 - 1) = 0.
  </math>

- '''Equation (2)''':
  <math>
  2x_2 + 2\lambda_1 x_2 + 2\lambda_2 x_2 = 0.
  </math>

From '''Equation (2)''', since <math>x_2 (1 + \lambda_1 + \lambda_2) = 0</math> and <math>\lambda_i \geq 0</math> for <math>i = 1, 2</math>, it follows that:

<math>
x_2 = 0.
</math>

Substitute <math>x_2 = 0</math> into the constraints:

<math>
\begin{aligned}
x_1^2 - 1 &\leq 0 \quad \Rightarrow \quad x_1 \in [-1, 1], \\
(x_1 - 1)^2 - 1 &\leq 0 \quad \Rightarrow \quad x_1 \in [0, 2].
\end{aligned}
</math>

'''Combining both constraints''':

<math>
x_1 \in [0, 1].
</math>

=== Step 4: Solve the problem Using Equation (1) ===

Substitute <math>x_2 = 0</math> into '''Equation (1)''':

<math>
(x_1 - 2) + \lambda_1 x_1 + \lambda_2 (x_1 - 1) = 0.
</math>

'''Assume <math>\lambda_1 > 0</math>''' (since Constraint 1 is active):

<math>
x_1^2 - 1 = 0 \quad \Rightarrow \quad x_1 = \pm 1.
</math>

But from the feasible range, <math>x_1 = 1</math>.

Substitute <math>x_1 = 1</math> into the equation:

<math>
\lambda_1 = 1.
</math>

This is acceptable.

'''Assume <math>\lambda_2 = 0</math>''' because Constraint 2 is not active at <math>x_1 = 1</math>.

=== Step 5: Verify Complementary Slackness ===

- '''Constraint 1''':
  
  <math>
  \lambda_1 (x_1^2 - 1) = 1 \times (1 - 1) = 0.
  </math>

- '''Constraint 2''':
  
  <math>
  \lambda_2 \left( (x_1 - 1)^2 + x_2^2 - 1 \right) = 0 \times (-1) = 0.
  </math>

=== Step 6: Verify Primal Feasibility ===

- '''Constraint 1''':
  
  <math>
  x_1^2 - 1 = 1 - 1 = 0 \leq 0.
  </math>

- '''Constraint 2''':
  
  <math>
  (x_1 - 1)^2 + x_2^2 - 1 = -1 \leq 0.
  </math>

=== Step 7: Conclusion ===

- '''Optimal Solution''':
  <math>
  x_1^* = 1, \quad x_2^* = 0.
  </math>

- '''Minimum Objective Value''':
  <math>
  f_0^*(x) = (1 - 2)^2 + 0 = 1.
  </math>

== Numerical Example 2 (SDP-Based QCQP) ==

Consider the following '''Quadratically Constrained Quadratic Programming (QCQP)''' problem:

<math> \begin{aligned} \text{minimize} \quad & f_0(x) = x_1^2 + x_2^2 \\ \text{subject to} \quad & f_1(x) = x_1^2 + x_2^2 - 2 \leq 0, \\ & f_2(x) = -x_1 x_2 + 1 \leq 0. \end{aligned} </math>
Interpretation:

The objective <math>f_0(x) = x_1^2 + x_2^2</math> is the squared distance from the origin.A point is found in the feasible region that is as close as possible to the origin.

The constraint <math>f_1(x) = x_1^2 + x_2^2 - 2 \leq 0</math> restricts <math>(x_1, x_2)</math> to lie inside or on a circle of radius <math>\sqrt{2}</math>.
The constraint <math>f_2(x) = -x_1 x_2 + 1 \leq 0 \implies x_1 x_2 \geq 1</math> defines a hyperbolic region. To satisfy <math>x_1 x_2 \geq 1</math>, both variables must be sufficiently large in magnitude and have the same sign.
=== Step 1: Lifting and Reformulation ===

Introduce the lifted variable:

<math> x = \begin{pmatrix} x_1 \\ x_2 \end{pmatrix}, \quad X = x x^T = \begin{pmatrix} x_1^2 & x_1 x_2 \\ x_1 x_2 & x_2^2 \end{pmatrix}. </math>
If <math>X = x x^T</math>, then <math>X \succeq 0</math> (positive semidefinite) and <math>X</math> is rank-1.

Rewrite the objective and constraints in terms of <math>X</math>:

Objective: <math>x_1^2 + x_2^2 = \langle I, X \rangle</math>, where <math>I</math> is the 2x2 identity matrix.

Constraint 1: <math>x_1^2 + x_2^2 - 2 \leq 0 \implies \langle I, X \rangle - 2 \leq 0.</math>

Constraint 2: <math>-x_1 x_2 + 1 \leq 0 \implies X_{12} \geq 1.</math>

=== Step 2: SDP Relaxation ===

The original QCQP is non-convex due to the rank-1 condition on <math>X</math>. Relax the rank constraint and consider only <math>X \succeq 0</math>:

<math> \begin{aligned} \text{minimize} \quad & \langle I, X \rangle \\ \text{subject to} \quad & \langle I, X \rangle - 2 \leq 0, \\ & X_{12} \geq 1, \\ & X \succeq 0. \end{aligned} </math>
This is now a Semidefinite Program (SDP), which is convex and can be solved using standard SDP solvers.

=== Step 3: Solving the SDP and Recovering the Solution ===

Solving the SDP, the feasible solution <math>X^*</math> found that achieves the minimum:

<math> X^* = \begin{pmatrix} 1 & 1 \\ 1 & 1 \end{pmatrix}. </math>
Check that <math>X^*</math> is rank-1:

<math> X^* = \begin{pmatrix}1 \\ 1\end{pmatrix} \begin{pmatrix}1 & 1\end{pmatrix} = x^*(x^*)^T, </math>
with <math>x^* = (1, 1)</math>.

Thus, from <math> \text{Thus, from } X^*, \text{ we recover the solution } x^* = (1,1) \text{ for the original QCQP.} </math>

Check feasibility:

<math>x_1^2 + x_2^2 = 1 + 1 = 2 \implies f_1(x^*) = 0 \leq 0.</math>
<math>x_1 x_2 = 1 \implies f_2(x^*) = -1 + 1 = 0 \leq 0.</math>
All constraints are satisfied.

=== Step 4: Optimal Value ===

The optimal objective value is: <math>f_0^*(x) = x_1^{*2} + x_2^{*2} = 1 + 1 = 2.</math>

=== Conclusion ===

The SDP relaxation not only provides a lower bound but also recovers the global optimum for this particular QCQP. The optimal solution is <math>(x_1^, x_2^) = (1, 1)</math> with an objective value of <math>2</math>.

This example demonstrates how an SDP relaxation can be used effectively to solve a non-convex QCQP and, in some cases, recover the exact optimal solution.

==Application==


==Conclusion==
In conclusion, Quadratically Constrained Quadratic Programs (QCQPs) are a significant class of optimization problems extending quadratic programming by incorporating quadratic constraints (Bao,Sahinidis,2011). These problems are essential for modeling complex real-world scenarios where both the objective function and the constraints are quadratic. QCQPs are widely applicable in areas such as agriculture, finance, production planning, and machine learning, where they help optimize decisions by balancing competing factors such as profitability, risk, and resource constraints.

The study and solution of QCQPs are critical due to their ability to capture complex relationships and non-linearities, offering a more realistic representation of many practical problems than simpler linear models. Techniques such as '''Karush-Kuhn Tucker (KKT) conditions, semidefinite programming (SDP) relaxations, and Reformulation-Linearization Technique (RLT)''' provide effective tools for solving QCQPs(Elloumi & Lambert,2019), offering both exact and approximate solutions depending on the problem’s structure. These methods allow for efficient handling of the challenges posed by quadratic constraints and non-linearities.

Looking forward, there are several potential areas for improvement in QCQP algorithms. One direction is the development of more efficient relaxation techniques for solving non-convex QCQPs, especially in large-scale problems where computational efficiency becomes critical. Additionally, there is ongoing research into hybrid methods that combine the strengths of different optimization techniques, such as SDP and machine learning, to improve the robustness and speed of solving QCQPs in dynamic environments. As optimization problems become increasingly complex and data-rich, advancements in QCQP algorithms will continue to play a crucial role in making informed, optimal decisions in diverse applications.
==Reference==
[1] Agarwal, D., Singh, P., & El Sayed, M. A. (2023). The Karush–Kuhn–Tucker (KKT) optimality conditions for fuzzy-valued fractional optimization problems. Mathematics and Computers in Simulation, 205, 861-877, DOI:[https://www.researchgate.net/publication/365290227_The_Karush-Kuhn-Tucker_KKT_optimality_conditions_for_fuzzy-valued_fractional_optimization_problems 10.1016/j.matcom.2022.10.024]

[2] Bao, X., Sahinidis, N. V., & Tawarmalani, M. (2011). [https://link.springer.com/article/10.1007/s10107-011-0462-2 Semidefinite relaxations for quadratically constrained quadratic programming: A review and comparisons] (PDF). ''Mathematical programming'', ''129'', 129-157.

[3] Bose, S., Gayme, D. F., Chandy, K. M., & Low, S. H. (2015). Quadratically constrained quadratic programs on acyclic graphs with application to power flow. IEEE Transactions on Control of Network Systems, 2(3), 278-287,[https://ieeexplore.ieee.org/document/7035094 DOI: 10.1109/TCNS.2015.2401172]

[4] Elloumi, S., & Lambert, A. (2019). Global solution of non-convex quadratically constrained quadratic programs. ''Optimization methods and software'', ''34''(1), 98-114,doi: https://doi.org/10.1080/10556788.2017.1350675

[5] Freund, R. M. (2004). [https://ocw.mit.edu/courses/15-084j-nonlinear-programming-spring-2004/a632b565602fd2eb3be574c537eea095_lec23_semidef_opt.pdf Introduction to semidefinite programming (SDP)] (PDF). Massachusetts Institute of Technology, 8-11.

[6] Ghojogh, B., Ghodsi, A., Karray, F., & Crowley, M. (2021). KKT conditions, first-order and second-order optimization, and distributed optimization: tutorial and survey. arXiv preprint arXiv:2110.01858.

[5] Zenios, S. A. (Ed.). (1993). Financial optimization. Cambridge university press,doi:https://doi.org/10.1017/CBO9780511522130

==Further Reading==
In Algorithm:

* [1] Floudas, C. A., & Visweswaran, V. (1995). [https://link.springer.com/chapter/10.1007/978-1-4615-2025-2_5 Quadratic optimization. ''Handbook of global optimization''] ''(PDF)'', 217-269.
* [2] Xu, H. K. (2003). [https://link.springer.com/article/10.1023/A:1023073621589 An iterative approach to quadratic optimization. ''Journal of Optimization Theory and Applications'']''(PDF)'', ''116'', 659-678.

In Finance Portfolio Construction:

* [1] [https://link.springer.com/article/10.1007/s10462-022-10273-7 Gunjan, A., & Bhattacharyya, S. (2023). A brief review of portfolio optimization techniques](PDF). ''Artificial Intelligence Review'', ''56''(5), 3847-3886,doi: https://doi.org/10.1007/s10462-022-10273-7
* [2] Xu, H. K. (2003). An iterative approach to quadratic optimization. ''Journal of Optimization Theory and Applications'', ''116'', 659-678.

==External Links==

* [https://nag.com/solving-quadratically-constrained-quadratic-programming-qcqp-problems/ Solving quadratically constrained quadratic programming (QCQP) problems]
{| class="wikitable"
|+
!Category: '''[[NonLinear Programming (NLP)]] - [[Quadratic programming]]'''
|}
<references />