Namespaces
Variants
Actions

Difference between revisions of "Ekeland variational principle"

From Encyclopedia of Mathematics
Jump to: navigation, search
(Importing text file)
 
m (tex encoded by computer)
 
(One intermediate revision by one other user not shown)
Line 1: Line 1:
There are usually three ways for getting existence results in analysis, namely compactness, Hahn–Banach-type results and completeness properties (cf. [[Compactness|Compactness]]; [[Hahn–Banach theorem|Hahn–Banach theorem]]; [[Completeness (in topology)|Completeness (in topology)]]). The Ekeland variational principle [[#References|[a10]]] (which provides a characterization of complete metric spaces [[#References|[a14]]], cf. also [[Complete metric space|Complete metric space]]) illustrates the third method in the framework of optimization. Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e1100301.png" /> be a lower [[Semi-continuous function|semi-continuous function]] defined on a [[Complete metric space|complete metric space]] <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e1100302.png" />, with values in the extended line <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e1100303.png" />, and bounded from below. It is well known that the lower bound of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e1100304.png" /> over <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e1100305.png" /> need not be attained. Ekeland's basic principle asserts that there exists a slight perturbation of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e1100306.png" /> which attains its minimum on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e1100307.png" />. More precisely, there exists a point <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e1100308.png" /> such that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e1100309.png" /> for all <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003010.png" />; this says that the function <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003011.png" /> has a strict minimum on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003012.png" /> at <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003013.png" />. It is interesting to observe that the conclusion of the basic principle is equivalent to the existence of a maximal element in the epigraph <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003014.png" /> for the order defined on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003015.png" /> by <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003016.png" /> if and only if <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003017.png" /> [[#References|[a3]]].
+
<!--
 +
e1100301.png
 +
$#A+1 = 63 n = 1
 +
$#C+1 = 63 : ~/encyclopedia/old_files/data/E110/E.1100030 Ekeland variational principle
 +
Automatically converted into TeX, above some diagnostics.
 +
Please remove this comment and the {{TEX|auto}} line below,
 +
if TeX found to be correct.
 +
-->
 +
 
 +
{{TEX|auto}}
 +
{{TEX|done}}
 +
 
 +
There are usually three ways for getting existence results in analysis, namely compactness, Hahn–Banach-type results and completeness properties (cf. [[Compactness|Compactness]]; [[Hahn–Banach theorem|Hahn–Banach theorem]]; [[Completeness (in topology)|Completeness (in topology)]]). The Ekeland variational principle [[#References|[a10]]] (which provides a characterization of complete metric spaces [[#References|[a14]]], cf. also [[Complete metric space|Complete metric space]]) illustrates the third method in the framework of optimization. Let $  f $
 +
be a lower [[Semi-continuous function|semi-continuous function]] defined on a [[Complete metric space|complete metric space]] $  ( X,d ) $,  
 +
with values in the extended line $  \mathbf R \cup \{ + \infty \} $,  
 +
and bounded from below. It is well known that the lower bound of $  f $
 +
over $  X $
 +
need not be attained. Ekeland's basic principle asserts that there exists a slight perturbation of $  f $
 +
which attains its minimum on $  X $.  
 +
More precisely, there exists a point $  a \in X $
 +
such that $  f ( a ) < f ( x ) + d ( a,x ) $
 +
for all $  x \in X \setminus  \{ a \} $;  
 +
this says that the function $  f ( \cdot ) + d ( a, \cdot ) $
 +
has a strict minimum on $  X $
 +
at $  a $.  
 +
It is interesting to observe that the conclusion of the basic principle is equivalent to the existence of a maximal element in the [[epigraph]]  $  { \mathop{\rm epi} } f = \{ {( x,r ) \in X \times \mathbf R } : {f ( x ) \leq  r } \} $
 +
for the order defined on $  X \times \mathbf R $
 +
by $  ( x _ {1} ,r _ {1} ) \cle ( x _ {2} ,r _ {2} ) $
 +
if and only if $  r _ {2} - r _ {1} + d ( x _ {2} ,x _ {1} ) \leq  0 $[[#References|[a3]]].
  
 
<img style="border:1px solid;" src="https://www.encyclopediaofmath.org/legacyimages/common_img/e110030a.gif" />
 
<img style="border:1px solid;" src="https://www.encyclopediaofmath.org/legacyimages/common_img/e110030a.gif" />
Line 5: Line 33:
 
Figure: e110030a
 
Figure: e110030a
  
From this basic principle one can deduce some variants which are in fact equivalent to the basic statement. The first one is as follows: given <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003018.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003019.png" /> such that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003020.png" /> and applying the basic principle to the complete metric space <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003021.png" />, one obtains the existence of a point <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003022.png" /> such that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003023.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003024.png" /> for all <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003025.png" />. In particular, this implies that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003026.png" />. Applying the previous result with the metric <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003027.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003028.png" />, yields the second variant: there exists an <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003029.png" /> such that
+
From this basic principle one can deduce some variants which are in fact equivalent to the basic statement. The first one is as follows: given $  \epsilon > 0 $,  
 +
$  x _ {0} \in X $
 +
such that $  f ( x _ {0} ) < \inf  _ {x \in X }  f ( x ) + \epsilon $
 +
and applying the basic principle to the complete metric space $  {\widetilde{X}  } = \{ {z \in X } : {f ( z ) + d ( x _ {0} ,z ) \leq  f ( x _ {0} ) } \} $,  
 +
one obtains the existence of a point $  a \in X $
 +
such that $  f ( a ) + d ( a,x _ {0} ) \leq  f ( x _ {0} ) $
 +
and $  f ( a ) < f ( x ) + d ( a,x ) $
 +
for all $  x \in X \setminus  \{ a \} $.  
 +
In particular, this implies that $  | {f ( a ) - f ( x _ {0} ) } | \leq  \epsilon $.  
 +
Applying the previous result with the metric $  \gamma d $,
 +
$  \gamma > 0 $,  
 +
yields the second variant: there exists an $  a \in X $
 +
such that
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003030.png" /></td> </tr></table>
+
$$
 +
d ( a,x _ {0} ) \leq  \gamma ^ {- 1 } \epsilon,
 +
$$
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003031.png" /></td> </tr></table>
+
$$
 +
\left | {f ( a ) - f ( x _ {0} ) } \right | \leq  \epsilon,
 +
$$
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003032.png" /></td> </tr></table>
+
$$
 +
f ( a ) < f ( x ) + \gamma d ( a,x )  \textrm{ for  all  }  x \in X \setminus  \{ a \} .
 +
$$
  
This variational principle has several equivalent geometric formulations. For instance, the Phelps extremization principle and the Drop theorem [[#References|[a7]]], [[#References|[a12]]] (see [[#References|[a13]]] for the versions as stated here). Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003033.png" /> be a closed subset of a [[Banach space|Banach space]] <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003034.png" />, let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003035.png" /> and let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003036.png" /> be a closed convex bounded subset of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003037.png" /> such that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003038.png" />. Then there exist a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003039.png" /> and a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003040.png" /> such that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003041.png" />.
+
This variational principle has several equivalent geometric formulations. For instance, the Phelps extremization principle and the Drop theorem [[#References|[a7]]], [[#References|[a12]]] (see [[#References|[a13]]] for the versions as stated here). Let $  A $
 +
be a closed subset of a [[Banach space|Banach space]] $  X $,  
 +
let $  x \in A $
 +
and let $  C $
 +
be a closed convex bounded subset of $  X $
 +
such that $  A \cap ( x + C ) = \emptyset $.  
 +
Then there exist a $  z \in A \cap ( x + [ 0,1 ] C ) $
 +
and a $  \delta > 0 $
 +
such that $  A \cap ( z + \left ] 0, \delta \right ] C ) = \emptyset $.
  
Among the great number of applications is the celebrated Bröndsted–Rockafellar theorem in [[Convex analysis|convex analysis]] [[#References|[a6]]]. Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003042.png" /> be a closed convex function defined on a real Banach space <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003043.png" /> with values in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003044.png" /> (cf. also [[Convex function (of a real variable)|Convex function (of a real variable)]]). Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003045.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003046.png" />, and let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003047.png" /> be such that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003048.png" /> for all <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003049.png" />. One can apply the third version of the theorem, with <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003050.png" />, to the function <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003051.png" /> when endowing <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003052.png" /> with the equivalent norm <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003053.png" /> [[#References|[a4]]]. This yields the existence of an <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003054.png" /> and an
+
Among the great number of applications is the celebrated Bröndsted–Rockafellar theorem in [[Convex analysis|convex analysis]] [[#References|[a6]]]. Let $  f $
 +
be a closed convex function defined on a real Banach space $  ( X, \| \cdot \| ) $
 +
with values in $  \mathbf R \cup \{ + \infty \} $(
 +
cf. also [[Convex function (of a real variable)|Convex function (of a real variable)]]). Let $  x _ {0} \in { \mathop{\rm dom} } f = \{ {x \in X } : {f ( x ) < + \infty } \} $,
 +
$  \epsilon > 0 $,  
 +
and let $  {\mathcal l} _ {0} \in X  ^ {*} $
 +
be such that $  f ( x ) \geq  f ( x _ {0} ) + {\mathcal l} _ {0} ( x - x _ {0} ) - \epsilon $
 +
for all $  x \in X $.  
 +
One can apply the third version of the theorem, with $  \gamma = \sqrt \epsilon $,  
 +
to the function $  g = f - {\mathcal l} _ {0} $
 +
when endowing $  X $
 +
with the equivalent norm $  \| \cdot \| + | { {\mathcal l} ( \cdot ) } | $[[#References|[a4]]]. This yields the existence of an $  x _  \epsilon  \in X $
 +
and an
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003055.png" /></td> </tr></table>
+
$$
 +
{\mathcal l} _  \epsilon  \in \partial  f ( x _  \epsilon  ) =
 +
$$
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003056.png" /></td> </tr></table>
+
$$
 +
=  
 +
\left \{ { {\mathcal l} \in X  ^ {*} } : {f ( x ) \geq  f ( x _  \epsilon  ) + {\mathcal l} ( x - x _  \epsilon  )  \textrm{ for  all  }  x \in X } \right \}
 +
$$
  
 
such that
 
such that
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003057.png" /></td> </tr></table>
+
$$
 +
\left \| {x _  \epsilon  - x _ {0} } \right \| \leq  \sqrt \epsilon ,
 +
$$
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003058.png" /></td> </tr></table>
+
$$
 +
\left | {f ( x _  \epsilon  ) - f ( x _ {0} ) } \right | \leq  2 \epsilon + \sqrt \epsilon .
 +
$$
  
 
Hence the set
 
Hence the set
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003059.png" /></td> </tr></table>
+
$$
 +
{ \mathop{\rm dom} } \partial  f = \left \{ {x \in { \mathop{\rm dom} } f } : {\partial  f ( x ) \neq \emptyset } \right \}
 +
$$
  
is dense in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003060.png" /> for the epigraph topology, i.e. the supremum of the norm topology on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003061.png" /> and of the initial topology associated to <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003062.png" />.
+
is dense in $  { \mathop{\rm dom} } f $
 +
for the epigraph topology, i.e. the supremum of the norm topology on $  X $
 +
and of the initial topology associated to $  f $.
  
 
Another easy consequence of the Ekeland variational principle is a generalization to multi-functions of the Kirk–Caristi fixed-point theorem [[#References|[a2]]].
 
Another easy consequence of the Ekeland variational principle is a generalization to multi-functions of the Kirk–Caristi fixed-point theorem [[#References|[a2]]].
  
Finally, it should be mentioned that analogous results hold in Banach spaces with the perturbation <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003063.png" /> replaced by some smooth one [[#References|[a5]]].
+
Finally, it should be mentioned that analogous results hold in Banach spaces with the perturbation $  d ( a, \cdot ) $
 +
replaced by some smooth one [[#References|[a5]]].
  
 
====References====
 
====References====
 
<table><TR><TD valign="top">[a1]</TD> <TD valign="top">  H. Attouch,  H. Riahi,  "Stability results for the Ekeland's variational principle and cone extremal solutions"  ''Math. Oper. Res.'' , '''18'''  (1993)  pp. 173–201</TD></TR><TR><TD valign="top">[a2]</TD> <TD valign="top">  J.-P. Aubin,  I. Ekeland,  "Applied nonlinear analysis" , Wiley  (1984)</TD></TR><TR><TD valign="top">[a3]</TD> <TD valign="top">  E. Bishop,  R.R. Phelps,  "The support functional of a convex set"  P. Klee (ed.) , ''Convexity'' , ''Proc. Symp. Pure Math.'' , '''7''' , Amer. Math. Soc.  (1963)  pp. 27–35</TD></TR><TR><TD valign="top">[a4]</TD> <TD valign="top">  J.M. Borwein,  "A note on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003064.png" />-subgradients and maximal monotonicity"  ''Pacific J. Math.'' , '''103'''  (1982)  pp. 307–314</TD></TR><TR><TD valign="top">[a5]</TD> <TD valign="top">  J.M. Borwein,  R. Preiss,  "Smooth variational principle"  ''Trans. Amer. Math. Soc.'' , '''303'''  (1987)  pp. 517–527</TD></TR><TR><TD valign="top">[a6]</TD> <TD valign="top">  A. Bröndsted,  R.T. Rockafellar,  "On the subdifferentiability of convex functions"  ''Proc. Amer. Math. Soc.'' , '''16'''  (1965)  pp. 605–611</TD></TR><TR><TD valign="top">[a7]</TD> <TD valign="top">  F.H. Clarke,  "Optimization and nonsmooth analysis" , Wiley  (1983)</TD></TR><TR><TD valign="top">[a8]</TD> <TD valign="top">  J. Daneš,  "A geometric theorem useful in nonlinear functional analysis"  ''Boll. Un. Mat. Ital.'' , '''4'''  (1972)  pp. 369–375</TD></TR><TR><TD valign="top">[a9]</TD> <TD valign="top">  D.G. de Figueiredo,  "The Ekeland variational principle, tours and detours" , ''Lecture Notes Tata Inst.'' , Springer  (1989)</TD></TR><TR><TD valign="top">[a10]</TD> <TD valign="top">  I. Ekeland,  "On the variational principle"  ''J. Math. Anal. Appl.'' , '''47'''  (1974)  pp. 324–353</TD></TR><TR><TD valign="top">[a11]</TD> <TD valign="top">  I. Ekeland,  "Nonconvex minimization problems"  ''Bull. Amer. Math. Soc. (N.S.)'' , '''1'''  (1979)  pp. 443–474</TD></TR><TR><TD valign="top">[a12]</TD> <TD valign="top">  J.-P. Penot,  "The drop theorem, the petal theorem and Ekeland's variational principle"  ''Nonlinear Anal.: Theory, Methods, Appl.'' , '''10'''  (1986)  pp. 813–822</TD></TR><TR><TD valign="top">[a13]</TD> <TD valign="top">  J.S. Treiman,  "Characterization of Clarke's tangent and normal cones in finite and infinite dimensions"  ''Nonlinear Anal.: Theory, Methods, Appl.'' , '''7'''  (1983)  pp. 771–783</TD></TR><TR><TD valign="top">[a14]</TD> <TD valign="top">  J.D. Weston,  "A characterization of metric completeness"  ''Proc. Amer. Math. Soc.'' , '''64'''  (1977)  pp. 186–188</TD></TR></table>
 
<table><TR><TD valign="top">[a1]</TD> <TD valign="top">  H. Attouch,  H. Riahi,  "Stability results for the Ekeland's variational principle and cone extremal solutions"  ''Math. Oper. Res.'' , '''18'''  (1993)  pp. 173–201</TD></TR><TR><TD valign="top">[a2]</TD> <TD valign="top">  J.-P. Aubin,  I. Ekeland,  "Applied nonlinear analysis" , Wiley  (1984)</TD></TR><TR><TD valign="top">[a3]</TD> <TD valign="top">  E. Bishop,  R.R. Phelps,  "The support functional of a convex set"  P. Klee (ed.) , ''Convexity'' , ''Proc. Symp. Pure Math.'' , '''7''' , Amer. Math. Soc.  (1963)  pp. 27–35</TD></TR><TR><TD valign="top">[a4]</TD> <TD valign="top">  J.M. Borwein,  "A note on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e110/e110030/e11003064.png" />-subgradients and maximal monotonicity"  ''Pacific J. Math.'' , '''103'''  (1982)  pp. 307–314</TD></TR><TR><TD valign="top">[a5]</TD> <TD valign="top">  J.M. Borwein,  R. Preiss,  "Smooth variational principle"  ''Trans. Amer. Math. Soc.'' , '''303'''  (1987)  pp. 517–527</TD></TR><TR><TD valign="top">[a6]</TD> <TD valign="top">  A. Bröndsted,  R.T. Rockafellar,  "On the subdifferentiability of convex functions"  ''Proc. Amer. Math. Soc.'' , '''16'''  (1965)  pp. 605–611</TD></TR><TR><TD valign="top">[a7]</TD> <TD valign="top">  F.H. Clarke,  "Optimization and nonsmooth analysis" , Wiley  (1983)</TD></TR><TR><TD valign="top">[a8]</TD> <TD valign="top">  J. Daneš,  "A geometric theorem useful in nonlinear functional analysis"  ''Boll. Un. Mat. Ital.'' , '''4'''  (1972)  pp. 369–375</TD></TR><TR><TD valign="top">[a9]</TD> <TD valign="top">  D.G. de Figueiredo,  "The Ekeland variational principle, tours and detours" , ''Lecture Notes Tata Inst.'' , Springer  (1989)</TD></TR><TR><TD valign="top">[a10]</TD> <TD valign="top">  I. Ekeland,  "On the variational principle"  ''J. Math. Anal. Appl.'' , '''47'''  (1974)  pp. 324–353</TD></TR><TR><TD valign="top">[a11]</TD> <TD valign="top">  I. Ekeland,  "Nonconvex minimization problems"  ''Bull. Amer. Math. Soc. (N.S.)'' , '''1'''  (1979)  pp. 443–474</TD></TR><TR><TD valign="top">[a12]</TD> <TD valign="top">  J.-P. Penot,  "The drop theorem, the petal theorem and Ekeland's variational principle"  ''Nonlinear Anal.: Theory, Methods, Appl.'' , '''10'''  (1986)  pp. 813–822</TD></TR><TR><TD valign="top">[a13]</TD> <TD valign="top">  J.S. Treiman,  "Characterization of Clarke's tangent and normal cones in finite and infinite dimensions"  ''Nonlinear Anal.: Theory, Methods, Appl.'' , '''7'''  (1983)  pp. 771–783</TD></TR><TR><TD valign="top">[a14]</TD> <TD valign="top">  J.D. Weston,  "A characterization of metric completeness"  ''Proc. Amer. Math. Soc.'' , '''64'''  (1977)  pp. 186–188</TD></TR></table>

Latest revision as of 19:37, 5 June 2020


There are usually three ways for getting existence results in analysis, namely compactness, Hahn–Banach-type results and completeness properties (cf. Compactness; Hahn–Banach theorem; Completeness (in topology)). The Ekeland variational principle [a10] (which provides a characterization of complete metric spaces [a14], cf. also Complete metric space) illustrates the third method in the framework of optimization. Let $ f $ be a lower semi-continuous function defined on a complete metric space $ ( X,d ) $, with values in the extended line $ \mathbf R \cup \{ + \infty \} $, and bounded from below. It is well known that the lower bound of $ f $ over $ X $ need not be attained. Ekeland's basic principle asserts that there exists a slight perturbation of $ f $ which attains its minimum on $ X $. More precisely, there exists a point $ a \in X $ such that $ f ( a ) < f ( x ) + d ( a,x ) $ for all $ x \in X \setminus \{ a \} $; this says that the function $ f ( \cdot ) + d ( a, \cdot ) $ has a strict minimum on $ X $ at $ a $. It is interesting to observe that the conclusion of the basic principle is equivalent to the existence of a maximal element in the epigraph $ { \mathop{\rm epi} } f = \{ {( x,r ) \in X \times \mathbf R } : {f ( x ) \leq r } \} $ for the order defined on $ X \times \mathbf R $ by $ ( x _ {1} ,r _ {1} ) \cle ( x _ {2} ,r _ {2} ) $ if and only if $ r _ {2} - r _ {1} + d ( x _ {2} ,x _ {1} ) \leq 0 $[a3].

Figure: e110030a

From this basic principle one can deduce some variants which are in fact equivalent to the basic statement. The first one is as follows: given $ \epsilon > 0 $, $ x _ {0} \in X $ such that $ f ( x _ {0} ) < \inf _ {x \in X } f ( x ) + \epsilon $ and applying the basic principle to the complete metric space $ {\widetilde{X} } = \{ {z \in X } : {f ( z ) + d ( x _ {0} ,z ) \leq f ( x _ {0} ) } \} $, one obtains the existence of a point $ a \in X $ such that $ f ( a ) + d ( a,x _ {0} ) \leq f ( x _ {0} ) $ and $ f ( a ) < f ( x ) + d ( a,x ) $ for all $ x \in X \setminus \{ a \} $. In particular, this implies that $ | {f ( a ) - f ( x _ {0} ) } | \leq \epsilon $. Applying the previous result with the metric $ \gamma d $, $ \gamma > 0 $, yields the second variant: there exists an $ a \in X $ such that

$$ d ( a,x _ {0} ) \leq \gamma ^ {- 1 } \epsilon, $$

$$ \left | {f ( a ) - f ( x _ {0} ) } \right | \leq \epsilon, $$

$$ f ( a ) < f ( x ) + \gamma d ( a,x ) \textrm{ for all } x \in X \setminus \{ a \} . $$

This variational principle has several equivalent geometric formulations. For instance, the Phelps extremization principle and the Drop theorem [a7], [a12] (see [a13] for the versions as stated here). Let $ A $ be a closed subset of a Banach space $ X $, let $ x \in A $ and let $ C $ be a closed convex bounded subset of $ X $ such that $ A \cap ( x + C ) = \emptyset $. Then there exist a $ z \in A \cap ( x + [ 0,1 ] C ) $ and a $ \delta > 0 $ such that $ A \cap ( z + \left ] 0, \delta \right ] C ) = \emptyset $.

Among the great number of applications is the celebrated Bröndsted–Rockafellar theorem in convex analysis [a6]. Let $ f $ be a closed convex function defined on a real Banach space $ ( X, \| \cdot \| ) $ with values in $ \mathbf R \cup \{ + \infty \} $( cf. also Convex function (of a real variable)). Let $ x _ {0} \in { \mathop{\rm dom} } f = \{ {x \in X } : {f ( x ) < + \infty } \} $, $ \epsilon > 0 $, and let $ {\mathcal l} _ {0} \in X ^ {*} $ be such that $ f ( x ) \geq f ( x _ {0} ) + {\mathcal l} _ {0} ( x - x _ {0} ) - \epsilon $ for all $ x \in X $. One can apply the third version of the theorem, with $ \gamma = \sqrt \epsilon $, to the function $ g = f - {\mathcal l} _ {0} $ when endowing $ X $ with the equivalent norm $ \| \cdot \| + | { {\mathcal l} ( \cdot ) } | $[a4]. This yields the existence of an $ x _ \epsilon \in X $ and an

$$ {\mathcal l} _ \epsilon \in \partial f ( x _ \epsilon ) = $$

$$ = \left \{ { {\mathcal l} \in X ^ {*} } : {f ( x ) \geq f ( x _ \epsilon ) + {\mathcal l} ( x - x _ \epsilon ) \textrm{ for all } x \in X } \right \} $$

such that

$$ \left \| {x _ \epsilon - x _ {0} } \right \| \leq \sqrt \epsilon , $$

$$ \left | {f ( x _ \epsilon ) - f ( x _ {0} ) } \right | \leq 2 \epsilon + \sqrt \epsilon . $$

Hence the set

$$ { \mathop{\rm dom} } \partial f = \left \{ {x \in { \mathop{\rm dom} } f } : {\partial f ( x ) \neq \emptyset } \right \} $$

is dense in $ { \mathop{\rm dom} } f $ for the epigraph topology, i.e. the supremum of the norm topology on $ X $ and of the initial topology associated to $ f $.

Another easy consequence of the Ekeland variational principle is a generalization to multi-functions of the Kirk–Caristi fixed-point theorem [a2].

Finally, it should be mentioned that analogous results hold in Banach spaces with the perturbation $ d ( a, \cdot ) $ replaced by some smooth one [a5].

References

[a1] H. Attouch, H. Riahi, "Stability results for the Ekeland's variational principle and cone extremal solutions" Math. Oper. Res. , 18 (1993) pp. 173–201
[a2] J.-P. Aubin, I. Ekeland, "Applied nonlinear analysis" , Wiley (1984)
[a3] E. Bishop, R.R. Phelps, "The support functional of a convex set" P. Klee (ed.) , Convexity , Proc. Symp. Pure Math. , 7 , Amer. Math. Soc. (1963) pp. 27–35
[a4] J.M. Borwein, "A note on -subgradients and maximal monotonicity" Pacific J. Math. , 103 (1982) pp. 307–314
[a5] J.M. Borwein, R. Preiss, "Smooth variational principle" Trans. Amer. Math. Soc. , 303 (1987) pp. 517–527
[a6] A. Bröndsted, R.T. Rockafellar, "On the subdifferentiability of convex functions" Proc. Amer. Math. Soc. , 16 (1965) pp. 605–611
[a7] F.H. Clarke, "Optimization and nonsmooth analysis" , Wiley (1983)
[a8] J. Daneš, "A geometric theorem useful in nonlinear functional analysis" Boll. Un. Mat. Ital. , 4 (1972) pp. 369–375
[a9] D.G. de Figueiredo, "The Ekeland variational principle, tours and detours" , Lecture Notes Tata Inst. , Springer (1989)
[a10] I. Ekeland, "On the variational principle" J. Math. Anal. Appl. , 47 (1974) pp. 324–353
[a11] I. Ekeland, "Nonconvex minimization problems" Bull. Amer. Math. Soc. (N.S.) , 1 (1979) pp. 443–474
[a12] J.-P. Penot, "The drop theorem, the petal theorem and Ekeland's variational principle" Nonlinear Anal.: Theory, Methods, Appl. , 10 (1986) pp. 813–822
[a13] J.S. Treiman, "Characterization of Clarke's tangent and normal cones in finite and infinite dimensions" Nonlinear Anal.: Theory, Methods, Appl. , 7 (1983) pp. 771–783
[a14] J.D. Weston, "A characterization of metric completeness" Proc. Amer. Math. Soc. , 64 (1977) pp. 186–188
How to Cite This Entry:
Ekeland variational principle. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Ekeland_variational_principle&oldid=17233
This article was adapted from an original article by H. AttouchD. Azé (originator), which appeared in Encyclopedia of Mathematics - ISBN 1402006098. See original article