Convex optimization

Convex optimization is a subfield of mathematical optimization that studies the problem of minimizing convex functions over convex sets. Many classes of convex optimization problems admit polynomial-time algorithms,[1] whereas mathematical optimization is in general NP-hard.[2][3][4]

Convex optimization has applications in a wide range of disciplines, such as automatic control systems, estimation and signal processing, communications and networks, electronic circuit design,[5] data analysis and modeling, finance, statistics (optimal experimental design),[6] and structural optimization, where the approximation concept has proven to be efficient.[7][8] With recent advancements in computing and optimization algorithms, convex programming is nearly as straightforward as linear programming.[9]


A convex optimization problem is an optimization problem in which the objective function is a convex function and the feasible set is a convex set. A function   mapping some subset of  into   is convex if its domain is convex and for all   and all   in its domain, the following condition holds:  . A set S is convex if for all members   and all  , we have that  .

Concretely, a convex optimization problem is the problem of finding some   attaining


where the objective function   is convex, as is the feasible set  .[10][11] If such a point exists, it is referred to as an optimal point or solution; the set of all optimal points is called the optimal set. If   is unbounded below over   or the infimum is not attained, then the optimization problem is said to be unbounded. Otherwise, if   is the empty set, then the problem is said to be infeasible.[12]

Standard formEdit

A convex optimization problem is in standard form if it is written as


where   is the optimization variable, the function   is convex,  ,  , are convex, and  ,  , are affine.[12] This notation describes the problem of finding   that minimizes   among all   satisfying  ,   and  ,  . The function   is the objective function of the problem, and the functions   and   are the constraint functions.

The feasible set   of the optimization problem consists of all points   satisfying the constraints. This set is convex because   is convex, the sublevel sets of convex functions are convex, affine sets are convex, and the intersection of convex sets is convex.[13]

A solution to a convex optimization problem is any point   attaining  . In general, a convex optimization problem may have zero, one, or many solutions.

Many optimization problems can be equivalently formulated in this standard form. For example, the problem of maximizing a concave function   can be re-formulated equivalently as the problem of minimizing the convex function  . The problem of maximizing a concave function over a convex set is commonly called a convex optimization problem.


The following are useful properties of convex optimization problems:[14][12]

  • every local minimum is a global minimum;
  • the optimal set is convex;
  • if the objective function is strictly convex, then the problem has at most one optimal point.

These results are used by the theory of convex minimization along with geometric notions from functional analysis (in Hilbert spaces) such as the Hilbert projection theorem, the separating hyperplane theorem, and Farkas' lemma.

Uncertainty AnalysisEdit

Ben-Hain and Elishakoff[15] (1990), Elishakoff et al.[16] (1994) applied convex analysis to model uncertainty.


The following problem classes are all convex optimization problems, or can be reduced to convex optimization problems via simple transformations:[12][17]

A hierarchy of convex optimization problems. (LP: linear program, QP: quadratic program, SOCP second-order cone program, SDP: semidefinite program, CP: cone program.)

Lagrange multipliersEdit

Consider a convex minimization problem given in standard form by a cost function   and inequality constraints   for  . Then the domain   is:


The Lagrangian function for the problem is


For each point   in   that minimizes   over  , there exist real numbers   called Lagrange multipliers, that satisfy these conditions simultaneously:

  1.   minimizes   over all  
  2.   with at least one  
  3.   (complementary slackness).

If there exists a "strictly feasible point", that is, a point   satisfying


then the statement above can be strengthened to require that  .

Conversely, if some   in   satisfies (1)–(3) for scalars   with   then   is certain to minimize   over  .


Convex optimization problems can be solved by the following contemporary methods:[18]

Subgradient methods can be implemented simply and so are widely used.[21] Dual subgradient methods are subgradient methods applied to a dual problem. The drift-plus-penalty method is similar to the dual subgradient method, but takes a time average of the primal variables.


Extensions of convex optimization include the optimization of biconvex, pseudo-convex, and quasiconvex functions. Extensions of the theory of convex analysis and iterative methods for approximately solving non-convex minimization problems occur in the field of generalized convexity, also known as abstract convex analysis.

See alsoEdit


  1. ^ a b Nesterov & Nemirovskii 1994
  2. ^ Murty, Katta; Kabadi, Santosh (1987). "Some NP-complete problems in quadratic and nonlinear programming". Mathematical Programming. 39 (2): 117–129. doi:10.1007/BF02592948.
  3. ^ Sahni, S. "Computationally related problems," in SIAM Journal on Computing, 3, 262--279, 1974.
  4. ^ Quadratic programming with one negative eigenvalue is NP-hard, Panos M. Pardalos and Stephen A. Vavasis in Journal of Global Optimization, Volume 1, Number 1, 1991, pg.15-22.
  5. ^ Boyd & Vandenberghe 2004, p. 17
  6. ^ Chritensen/Klarbring, chpt. 4.
  7. ^ Boyd & Vandenberghe 2004
  8. ^ Schmit, L.A.; Fleury, C. 1980: Structural synthesis by combining approximation concepts and dual methods. J. Amer. Inst. Aeronaut. Astronaut 18, 1252-1260
  9. ^ Boyd & Vandenberghe 2004, p. 8
  10. ^ Hiriart-Urruty, Jean-Baptiste; Lemaréchal, Claude (1996). Convex analysis and minimization algorithms: Fundamentals. p. 291. ISBN 9783540568506.
  11. ^ Ben-Tal, Aharon; Nemirovskiĭ, Arkadiĭ Semenovich (2001). Lectures on modern convex optimization: analysis, algorithms, and engineering applications. pp. 335–336. ISBN 9780898714913.
  12. ^ a b c d Boyd & Vandenberghe 2004, chpt. 4
  13. ^ Boyd & Vandenberghe 2004, chpt. 2
  14. ^ Rockafellar, R. Tyrrell (1993). "Lagrange multipliers and optimality" (PDF). SIAM Review. 35 (2): 183–238. CiteSeerX doi:10.1137/1035044.
  15. ^ Ben Haim Y. and Elishakoff I., Convex Models of Uncertainty in Applied Mechanics, Elsevier Science Publishers, Amsterdam, 1990
  16. ^ I. Elishakoff, I. Lin Y.K. and Zhu L.P., Probabilistic and Convex Modeling of Acoustically Excited Structures, Elsevier Science Publishers, Amsterdam, 1994
  17. ^ Agrawal, Akshay; Verschueren, Robin; Diamond, Steven; Boyd, Stephen (2018). "A rewriting system for convex optimization problems" (PDF). Control and Decision. 5 (1): 42–60. arXiv:1709.04494. doi:10.1080/23307706.2017.1397554.
  18. ^ For methods for convex minimization, see the volumes by Hiriart-Urruty and Lemaréchal (bundle) and the textbooks by Ruszczyński, Bertsekas, and Boyd and Vandenberghe (interior point).
  19. ^ Nesterov, Yurii; Arkadii, Nemirovskii (1995). Interior-Point Polynomial Algorithms in Convex Programming. Society for Industrial and Applied Mathematics. ISBN 978-0898715156.
  20. ^ Peng, Jiming; Roos, Cornelis; Terlaky, Tamás (2002). "Self-regular functions and new search directions for linear and semidefinite optimization". Mathematical Programming. 93 (1): 129–171. doi:10.1007/s101070200296. ISSN 0025-5610.
  21. ^ Bertsekas


  • Ruszczyński, Andrzej (2006). Nonlinear Optimization. Princeton University Press.
  • Schmit, L.A.; Fleury, C. 1980: Structural synthesis by combining approximation concepts and dual methods. J. Amer. Inst. Aeronaut. Astronaut 18, 1252-1260

External linksEdit