Submodular Function Minimization Satoru Iwata (University of Tokyo)
Submodular Function Minimization ( ) 0 f Assumption: X Minimization Evaluation Algorithm Oracle ( X ) f Minimizer min{ ( ) | } ? f Y Y V
Submodular Function Minimization Grötschel, Lovász, Schrijver (1981, 1988) Ellipsoid Method 5 ( log ) Cunningham (1985) O n M 7 8 7 ( ) O n n ( log ) O n n Iwata, Fleischer, Fujishige (2000) Schrijver (2000) Fleischer, Iwata (2000) Iwata (2002) Fully Combinatorial Iwata (2003) Orlin (2007) 4 5 5 6 ( ) (( ) log ) O n n O n n M Iwata, Orlin (2009)
Submodular Functions and Polyhedra ( ) 0 V f : 2 f R ( ) ( ) ( ) ( ), , f X f Y f X Y f X Y X Y V Submodular Polyhedron V ( ) { | , , ( ) ( )} P f x x R Y V x Y f Y Base Polyhedron ( ) { | ( ), ( ) ( )} B f x x P f x V f V
Min-Max Theorem Edmonds (1970) Theorem min ( ) max{ ( ) | ( ), 0 } f Y z V z P f z Y V max{ ( ) | ( )} x V x B f ( ) : min{ 0 , ( )} x v x v ( ) ( ) ( ) z V z Y f Y ( ) ( ) ( ) x V x Y f Y
Combinatorial Approach min ( ) max{ ( ) | ( )} f Y x V x B f Y V
Min-Max Theorem min ( ) max{ ( ) | ( ), 0 } f Y z V z P f z Y V max{ ( ) | ( )} x V x B f ( ) : min{ ( ) | } f X f Y Y X Submodular ( ) ( ) P f P f ( ) ( ) ( ) min ( ) z B f z V f V f Y Y V
Combinatorial Approach min ( ) max{ ( ) | ( )} f Y x V x B f Y V Convex Combination x i y i i I y i ( f ) : Extreme Base Generated by B the Greedy Algorithm with an . Linear Ordering in L V i
IFF Scaling Algorithm 1 ( ) M f X Submodular : 2 2 n n ( ) ( ) | | | \ | Cut Function f X f X X V X : Flow in the Complete Digraph ( , ) u v x i y i - Augmenting Path i I s t x z S T z Increase ( ) V { | ( ) } { | ( ) } S v z v T v z v
IFF Scaling Algorithm Double-Exchange S T No Path from to : ( ) y y W i i u v : { | : Reachable from } v v S : min{ , } i W : ( ) x x u v ( , ) : ( , ) u v u v u v S T i i v u L i u v Saturating Nonsaturating
IFF Scaling Algorithm ( ) ( ), y i W f W i I T S No Path from to No Active Triple ( ) ( ) ( ) ( , , ) i u v x W y W f W i i i I ( ) ( ) z V f W n W 2 ( ) ( ) x V f W n S 1 T Min ( W ) : f 2 n L 5 ( log ) O n M i
Submodular Function Minimization Grötschel, Lovász, Schrijver (1981, 1988) Ellipsoid Method 5 ( log ) Cunningham (1985) O n M 7 8 7 ( ) O n n ( log ) O n n Iwata, Fleischer, Fujishige (2000) Schrijver (2000) Fleischer, Iwata (2000) Iwata (2002) Fully Combinatorial Iwata (2003) Orlin (2007) 4 5 5 6 ( ) (( ) log ) O n n O n n M Iwata, Orlin (2009)
Fully Combinatorial Algorithm Addition, Subtraction, Comparison Oracle Call (Function Evaluation) Compute , R Z Multiplication / q , R R Compute Division • Neglect the Gaussian Elimination Step. • Use Nonnegative Integer Combination Instead of Convex Combination. • Choose a Step Length Appropriately.
Finding All Minimizers , Y : , : X X Y X Y Minimizer Minimizer The set of minimizers forms a distributive lattice. [G.Birkhoff] Any distributive lattice can be represented as the set of ideals of a partial ordered set.
Partial Order of an Extreme Base y Extreme ( f ) : B Bixby, Cunningham, Topkis (1985) v v v ( ) ( ) : y X f X Tight X 2 n 1 Represent all tight sets ( ( )) ( ( )), y L v f L v v V ( y ) G v . u ( u ) : Maximal Ideal Excluding H u Test if is tight. ( ) { } H u v ( u ) H v u If not, then .
Finding All Minimizers Extreme Base ( ) ( x ) : G y G Superposition of i y i ( f ) B SCC Decomposition Partial Order (DAG) ( ) 0 x v ( ) G y i Convex Combination x ( ) 0 x v i y i i I ( 0 , ) i I i ( ) 0 x v
Recommend
More recommend