Fenchel-Moreau and Primal/Dual Optimization Problems: Difference between revisions

From Optimal Transport Wiki
Jump to navigation Jump to search
 
(33 intermediate revisions by the same user not shown)
Line 1: Line 1:
The Fenchel-Moreau Theorem is a fundamental result in convex analysis, characterizing the class of functions for which a function equals its biconjugate. A key consequence of this theorem is the equivalence of ''primal'' and ''dual'' optimization problems.
The Fenchel-Moreau Theorem<ref name="Brezis" /> is a fundamental result in convex analysis, characterizing the class of functions for which a function equals its biconjugate. A key consequence of this theorem is that is provides sufficient conditions for the equivalence of ''primal'' and ''dual'' optimization problems.<ref name="Rockafellar" />


==Background on Conjugate Functions==
==Fenchel-Moreau Theorem==
Let''X'' be a normed vector space, and let ''X*'' denote its topological dual. Given an extended real-valued function <math> f: X \to \mathbb{R} \cup \{+\infty\} </math>, its ''convex conjugate'' <math>f^*:X^* \to \mathbb{R} \cup \{+\infty\}</math> is defined by
Given a normed vector space ''X'' and <math> f: X \to \mathbb{R} \cup \{+\infty\} </math>, then


:<math> f^*(u):=\sup_{x \in X} \{ \langle u,x \rangle - f(x) \} \quad \forall u \in X^*. </math>
:<math> f \text{ is convex and lower semicontinuous} \iff f^{**} = f. </math>
 
==Background on Convex Conjugate Functions==
Let ''X'' be a normed vector space, and let ''X*'' denote its topological dual. Given an extended real-valued function <math> f: X \to \mathbb{R} \cup \{+\infty\} </math>, its ''convex conjugate'' <math>f^*:X^* \to \mathbb{R} \cup \{+\infty\}</math> is defined by
 
:<math> f^*(y):=\sup_{x \in X} \{ \langle y,x \rangle - f(x) \} \quad \forall y \in X^*. </math>


An immediate consequence of this definition is ''Young's Inequality'',
An immediate consequence of this definition is ''Young's Inequality'',


:<math> f^*(u) +f(x) \geq \langle u,x \rangle \quad \forall x \in X, u \in X^* . </math>
:<math> f^*(y) +f(x) \geq \langle y,x \rangle \quad \forall x \in X, y \in X^* . </math>


Furthermore, it follows directly from the definition that, for ''any'' function ''f'', its conjugate function ''f*'' is convex and lower semicontinuous.
Furthermore, it follows directly from the definition that, for ''any'' function ''f'', its conjugate function ''f*'' is convex and lower semicontinuous.


In a similar way, for any function ''f'', its the ''biconjugate'' function <math>f^{**}:X \to \mathbb{R} \cup \{+\infty\} <\math> is defined by  
In a similar way, for any function ''f'', its the ''biconjugate'' function <math>f^{**}:X \to \mathbb{R} \cup \{+\infty\} </math> is defined by
 
:<math> f^{**}(x):=\sup_{y \in X^*} \{ \langle y,x \rangle - f^*(y) \} \quad \forall y \in X^*. </math>
 
As above, the biconjugate function ''f**'' is always convex and lower semicontinuous. Furthermore, by a second application of Young's inequality, we have
 
:<math> f^{**}(x) \leq f(x) \quad \forall x \in X. </math>
 
Since ''f**'' is always convex and lower semicontinuous, in order for equality to hold for all ''x'', it is necessary that ''f'' also be convex and lower semicontinuous. The heart of Fenchel-Moreau Theorem is that this condition is not just necessary, but sufficient.
 
==Application to Primal/Dual Optimization Problems==
An important consequence of the Fenchel-Moreau Theorem is that it provides sufficient conditions for the equivalence of primal and dual optimization problems. Given a normed vector space ''X'' and a lower semicontinuous, convex function <math>f: X \to \mathbb{R}\cup \{+\infty\}</math>, the ''primal'' optimization problem is given by
 
:<math> \inf_{x \in X} f(x). </math>


:<math> f^{**}(x):=\sup_{u \in X^*} \{ \langle u,x \rangle - f^*(u) \} \quad \forall u \in X^*. </math>
The corresponding dual problem arises from a suitable ``perturbation" of the primal problem, subject to a parameter ''u'' ∈ ''U'', where ''U'' is also a normed vector space. In particular, let <math> F:X \times U \to \mathbb{R} \cup \{+\infty\} </math> be a proper convex function so that <math> f(x) = F(x,0) </math>. Then the corresponding ''primal'' and ''dual'' problems may be written as


As above, for any function ''f'', its biconjugate function ''f^*'' is convex and lower semicontinuous. Furthermore, by Young's inequality, we always have
:<math> P_0 := \inf_{x \in X} F(x,0) </math>
:<math> D_0 := \sup_{v \in U^*} -F^*(0,v) </math>


:<math> f^{**}(x) \leq f(x) \quad \forall x \in X. <\math>
The formulation of these problems becomes even simpler from the perspective of the ''inf-projection'' <math> P(u) := \inf_{x} F(x,u) </math>. With this notation, the primal and dual problems are given by


In this way, it is clear that, in order for equality to hold, it is necessary for ''f'' to be convex and lower semicontinuous. The heart of Fenchel-Moreau Theorem is that this is also sufficient.
:<math> P_0 = P(0)</math>
:<math> D_0 =P^{**}(0) </math>


Therefore, by the Fenchel-Moreau theorem, a sufficient condition for equivalence of the primal and dual problems is that the inf-projetion function ''P(u)'' is convex and lower semicontinuous.
<ref name="Brezis" />


==References==
==References==
<references>
<references>
<ref name="Brezis">H. Brezis, ''Functional Analysis''.</ref>
<ref name="Brezis">[https://link.springer.com/book/10.1007/978-0-387-70914-7 H. Brezis, ''Functional Analysis, Sobolev Spaces, and Partial Differential Equations'', Chapter 1.]</ref>
<ref name="Rockafellar">[https://link.springer.com/book/10.1007/978-3-642-02431-3 R. T. Rockafellar and R. J-B. Wets, ''Variational Analysis'', Chapter 11.]</ref>
</references>
</references>

Latest revision as of 04:39, 28 February 2022

The Fenchel-Moreau Theorem[1] is a fundamental result in convex analysis, characterizing the class of functions for which a function equals its biconjugate. A key consequence of this theorem is that is provides sufficient conditions for the equivalence of primal and dual optimization problems.[2]

Fenchel-Moreau Theorem

Given a normed vector space X and , then

Background on Convex Conjugate Functions

Let X be a normed vector space, and let X* denote its topological dual. Given an extended real-valued function , its convex conjugate is defined by

An immediate consequence of this definition is Young's Inequality,

Furthermore, it follows directly from the definition that, for any function f, its conjugate function f* is convex and lower semicontinuous.

In a similar way, for any function f, its the biconjugate function is defined by

As above, the biconjugate function f** is always convex and lower semicontinuous. Furthermore, by a second application of Young's inequality, we have

Since f** is always convex and lower semicontinuous, in order for equality to hold for all x, it is necessary that f also be convex and lower semicontinuous. The heart of Fenchel-Moreau Theorem is that this condition is not just necessary, but sufficient.

Application to Primal/Dual Optimization Problems

An important consequence of the Fenchel-Moreau Theorem is that it provides sufficient conditions for the equivalence of primal and dual optimization problems. Given a normed vector space X and a lower semicontinuous, convex function , the primal optimization problem is given by

The corresponding dual problem arises from a suitable ``perturbation" of the primal problem, subject to a parameter uU, where U is also a normed vector space. In particular, let be a proper convex function so that . Then the corresponding primal and dual problems may be written as

The formulation of these problems becomes even simpler from the perspective of the inf-projection . With this notation, the primal and dual problems are given by

Therefore, by the Fenchel-Moreau theorem, a sufficient condition for equivalence of the primal and dual problems is that the inf-projetion function P(u) is convex and lower semicontinuous.

References