projected chv atal gomory cuts for mixed integer linear
play

Projected Chv atal-Gomory cuts for Mixed Integer Linear Programs - PowerPoint PPT Presentation

Projected Chv atal-Gomory cuts for Mixed Integer Linear Programs Pierre Bonami CMU, USA Gerard Cornu ejols CMU, USA and LIF Marseille, France Sanjeeb Dash IBM T.J. Watson, USA Matteo Fischetti University of Padova, Italy Andrea Lodi


  1. Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs Pierre Bonami CMU, USA Gerard Cornu´ ejols CMU, USA and LIF Marseille, France Sanjeeb Dash IBM T.J. Watson, USA Matteo Fischetti University of Padova, Italy Andrea Lodi University of Bologna, Italy alodi@deis.unibo.it Aussois X, January 9, 2006 A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs

  2. Notation and background • Consider an Integer Linear Program (ILP) of the form: min { c T x : Ax ≤ b, x ≥ 0 integer } A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 1

  3. Notation and background • Consider an Integer Linear Program (ILP) of the form: min { c T x : Ax ≤ b, x ≥ 0 integer } and two associated polyhedra: R n P := { x ∈ I + : Ax ≤ b } P I := conv { x ∈ Z n + : Ax ≤ b } = conv ( P ∩ Z n ) A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 1

  4. Notation and background • Consider an Integer Linear Program (ILP) of the form: min { c T x : Ax ≤ b, x ≥ 0 integer } and two associated polyhedra: R n P := { x ∈ I + : Ax ≤ b } P I := conv { x ∈ Z n + : Ax ≤ b } = conv ( P ∩ Z n ) • A Chv´ atal-Gomory ( CG ) cut is a valid inequality for P I of the form: A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 1

  5. Notation and background • Consider an Integer Linear Program (ILP) of the form: min { c T x : Ax ≤ b, x ≥ 0 integer } and two associated polyhedra: R n P := { x ∈ I + : Ax ≤ b } P I := conv { x ∈ Z n + : Ax ≤ b } = conv ( P ∩ Z n ) • A Chv´ atal-Gomory ( CG ) cut is a valid inequality for P I of the form: ⌊ u T A ⌋ x ≤ ⌊ u T b ⌋ where u ∈ R m + is called the CG multiplier vector, and ⌊·⌋ denotes lower integer part. A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 1

  6. Notation and background • Consider an Integer Linear Program (ILP) of the form: min { c T x : Ax ≤ b, x ≥ 0 integer } and two associated polyhedra: R n P := { x ∈ I + : Ax ≤ b } P I := conv { x ∈ Z n + : Ax ≤ b } = conv ( P ∩ Z n ) • A Chv´ atal-Gomory ( CG ) cut is a valid inequality for P I of the form: ⌊ u T A ⌋ x ≤ ⌊ u T b ⌋ where u ∈ R m + is called the CG multiplier vector, and ⌊·⌋ denotes lower integer part. • The first Chv´ atal closure of P is defined as: P 1 := { x ≥ 0 : Ax ≤ b, ⌊ u T A ⌋ x ≤ ⌊ u T b ⌋ for all u ∈ I R m + } A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 1

  7. Notation and background • Consider an Integer Linear Program (ILP) of the form: min { c T x : Ax ≤ b, x ≥ 0 integer } and two associated polyhedra: R n P := { x ∈ I + : Ax ≤ b } P I := conv { x ∈ Z n + : Ax ≤ b } = conv ( P ∩ Z n ) • A Chv´ atal-Gomory ( CG ) cut is a valid inequality for P I of the form: ⌊ u T A ⌋ x ≤ ⌊ u T b ⌋ where u ∈ R m + is called the CG multiplier vector, and ⌊·⌋ denotes lower integer part. • The first Chv´ atal closure of P is defined as: P 1 := { x ≥ 0 : Ax ≤ b, ⌊ u T A ⌋ x ≤ ⌊ u T b ⌋ for all u ∈ I R m + } [Chv´ atal 1973] • P 1 is indeed a polyhedron, i.e., a finite number of CG cuts suffice to define it. A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 1

  8. Notation and background (cont.d) • Clearly, P I ⊆ P 1 ⊆ P . A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 2

  9. Notation and background (cont.d) • Clearly, P I ⊆ P 1 ⊆ P . • Chv´ atal-Gomory separation problem (CG-SEP) is NP-hard. [Eisenbrand 1999] A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 2

  10. Notation and background (cont.d) • Clearly, P I ⊆ P 1 ⊆ P . • Chv´ atal-Gomory separation problem (CG-SEP) is NP-hard. [Eisenbrand 1999] • Recently Fischetti & Lodi have shown that: Optimizing over P 1 is possible in practice via an MIP model (MIPping). A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 2

  11. Notation and background (cont.d) • Clearly, P I ⊆ P 1 ⊆ P . • Chv´ atal-Gomory separation problem (CG-SEP) is NP-hard. [Eisenbrand 1999] • Recently Fischetti & Lodi have shown that: Optimizing over P 1 is possible in practice via an MIP model (MIPping). P 1 is a good (often excellent) approximation of P I in practice. A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 2

  12. Notation and background (cont.d) • Clearly, P I ⊆ P 1 ⊆ P . • Chv´ atal-Gomory separation problem (CG-SEP) is NP-hard. [Eisenbrand 1999] • Recently Fischetti & Lodi have shown that: Optimizing over P 1 is possible in practice via an MIP model (MIPping). P 1 is a good (often excellent) approximation of P I in practice. CG cuts in the first closure have a nice numerical behavior and stability. A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 2

  13. Notation and background (cont.d) • Clearly, P I ⊆ P 1 ⊆ P . • Chv´ atal-Gomory separation problem (CG-SEP) is NP-hard. [Eisenbrand 1999] • Recently Fischetti & Lodi have shown that: Optimizing over P 1 is possible in practice via an MIP model (MIPping). P 1 is a good (often excellent) approximation of P I in practice. CG cuts in the first closure have a nice numerical behavior and stability. • Thus, the natural question is: What does it happen in the Mixed IP case? A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 2

  14. Notation and background (cont.d) • Clearly, P I ⊆ P 1 ⊆ P . • Chv´ atal-Gomory separation problem (CG-SEP) is NP-hard. [Eisenbrand 1999] • Recently Fischetti & Lodi have shown that: Optimizing over P 1 is possible in practice via an MIP model (MIPping). P 1 is a good (often excellent) approximation of P I in practice. CG cuts in the first closure have a nice numerical behavior and stability. • Thus, the natural question is: What does it happen in the Mixed IP case? • Of course, the natural answer would be using Gomory Mixed Integer cuts (GMI) (also known as MIR cuts and split cuts) but their separation is much more involved than CG separation: nobody knows a MIP model for GMI yet! A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 2

  15. Projected Chv´ atal-Gomory cuts • Our first order of business is to extend the classical definition of Chv´ atal-Gomory cuts to the mixed integer case, in such a way that the resulting separation problem remains a clean MIP. A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 3

  16. Projected Chv´ atal-Gomory cuts • Our first order of business is to extend the classical definition of Chv´ atal-Gomory cuts to the mixed integer case, in such a way that the resulting separation problem remains a clean MIP. • We then consider the MIP: min { c T x + f T y : Ax + Cy ≤ b, x ≥ 0 , x integral , y ≥ 0 } • with the two associated polyhedra: P ( x, y ) := { ( x, y ) ∈ R n + × R r + : Ax + Cy ≤ b } P I ( x, y ) := conv ( { ( x, y ) ∈ P ( x, y ) : x integral } ) A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 3

  17. Projected Chv´ atal-Gomory cuts • Our first order of business is to extend the classical definition of Chv´ atal-Gomory cuts to the mixed integer case, in such a way that the resulting separation problem remains a clean MIP. • We then consider the MIP: min { c T x + f T y : Ax + Cy ≤ b, x ≥ 0 , x integral , y ≥ 0 } • with the two associated polyhedra: P ( x, y ) := { ( x, y ) ∈ R n + × R r + : Ax + Cy ≤ b } P I ( x, y ) := conv ( { ( x, y ) ∈ P ( x, y ) : x integral } ) • and we project P ( x, y ) onto the space of x variables as: := { x ∈ R n + : there exists y ∈ R r P ( x ) + s.t. Ax + Cy ≤ b } A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 3

  18. Projected Chv´ atal-Gomory cuts • Our first order of business is to extend the classical definition of Chv´ atal-Gomory cuts to the mixed integer case, in such a way that the resulting separation problem remains a clean MIP. • We then consider the MIP: min { c T x + f T y : Ax + Cy ≤ b, x ≥ 0 , x integral , y ≥ 0 } • with the two associated polyhedra: P ( x, y ) := { ( x, y ) ∈ R n + × R r + : Ax + Cy ≤ b } P I ( x, y ) := conv ( { ( x, y ) ∈ P ( x, y ) : x integral } ) • and we project P ( x, y ) onto the space of x variables as: := { x ∈ R n + : there exists y ∈ R r P ( x ) + s.t. Ax + Cy ≤ b } = { x ∈ R n + : u k A ≤ u k b, k = 1 , . . . , K } A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 3

  19. Projected Chv´ atal-Gomory cuts • Our first order of business is to extend the classical definition of Chv´ atal-Gomory cuts to the mixed integer case, in such a way that the resulting separation problem remains a clean MIP. • We then consider the MIP: min { c T x + f T y : Ax + Cy ≤ b, x ≥ 0 , x integral , y ≥ 0 } • with the two associated polyhedra: P ( x, y ) := { ( x, y ) ∈ R n + × R r + : Ax + Cy ≤ b } P I ( x, y ) := conv ( { ( x, y ) ∈ P ( x, y ) : x integral } ) • and we project P ( x, y ) onto the space of x variables as: := { x ∈ R n + : there exists y ∈ R r P ( x ) + s.t. Ax + Cy ≤ b } = { x ∈ R n + : u k A ≤ u k b, k = 1 , . . . , K } =: { x ∈ R n + : ¯ Ax ≤ ¯ b } where u 1 , . . . , u K are the (finitely many) extreme rays of the projection cone { u ∈ R m + : u T C ≥ 0 T } . A. Lodi, Projected Chv´ atal-Gomory cuts for Mixed Integer Linear Programs 3

Recommend


More recommend