DUALITY AND INTEGER PROGRAMMING. Jean B. LASSERRE

Size: px
Start display at page:

Download "DUALITY AND INTEGER PROGRAMMING. Jean B. LASSERRE"

Transcription

1 LABORATOIRE d ANALYSE et d ARCHITECTURE des SYSTEMES DUALITY AND INTEGER PROGRAMMING Jean B. LASSERRE 1

2 Current solvers (CPLEX, XPRESS-MP) are rather efficient and can solve many large size problems with thousands of variables. However, the very small 5-variables knapsack problem max 213x x x x x x x x x x 5 = x 1, x 2, x 3, x 4, x 5 nonnegative integers still resists all efficient solvers (takes HOURS on CPLEX 8.1 and XPRESS-MP!) Optimal solution x = (7334, 0, 0, 0, 0)...! Any insight on integer problems is welcome... 2

3 The integer program max{ c x Ax = b; x N n } with A = [A 1,..., A n ] Z m n, b Z m has a dual problem min { f(b) f(a j) c j, j = 1,..., n} f Γ where Γ is the space of functions f : R m R that are superadditive (f(a + b) f(a) + f(b)), and with f(0) = 0. (Jeroslow, Wolsey, etc...) Elegant but rather abstract and not very practical; however, used to derive valid inequalities. Moreover, the celebrated Gomory cuts used in Solvers (CPLEX, XPRESS-MP,..) can be interpreted as such superadditive functions. 3

4 CONTINUOUS OPTIM. DISCRETE OPTIM. f(b, c) := max c x f d (b, c) := max c x [ [ Ax b Ax b s.t. x R n s.t. + x N n INTEGRATION SUMMATION f(b, c) := Ω ec x dx f d (b, c) := e c x [ [ Ω Ax b Ax b Ω := x R n Ω := x N n + 4

5 e f(b,c) = lim f(b, rc) 1/r ; e fd(b,c) = lim f r r d (b, rc) 1/r. or, equivalently 1 f(b, c) = lim r r ln f(b, 1 rc); f d (b, c) = lim r r ln f d (b, rc). Note in passing that we have the mean-value theorem f(b, c) = e c x λ(ω) = e c x vol(ω). and a Maslov max mean-value theorem f(b, c) = c x = c x µ maslov (Ω) = c x. Maslov duality for linear functionals in the Max-Plus algebra 5

6 f L(f) a linear functional in the max + algebra, i.e., L(f h) = L(f) L(h) f, h L(f) = with f dµ maslov. µ maslov (A B) = µ maslov (A) µ maslov (B) A, B B. 6

7 CONTINUOUS OPTIM. DISCRETE OPTIM. Legendre-Fenchel Duality?? INTEGRATION SUMMATION Laplace-Transform Z-Transform Duality Duality 7

8 Legendre-Fenchel duality : f : R n R convex; f : R n R. λ f (λ) = F(f)(λ) := sup y {λ y f(y)}. (One-sided) Laplace-Transform: f : R n + R; F : Cn C. λ F (λ) = L(f)(λ) := R n + e λ y f(y) dy. (One-sided) Z-Transform: f : Z n + R; F : Cn C. λ F (z) = Z(f)(z) := m Z n + z m f(m). 8

9 Observe : and e F(f)(λ) = L(e f )(λ) = R n + sup (λ y f(y)) e y e (λ y f(y)) dy = sup y {e (λ y f(y)) } Maslov e(λ y f(y)) dy exponential (Fenchel (f)) = Laplace (exponential (f)) in max, algebra 9

10 with Ω := {x R n Ax b; x 0} Fenchel-duality: Laplace-duality f(b, c) := max x Ω c x f(b, c) := f (λ, c) := min y {λ y f(y, c)} F (λ, c) := = Ω ec x dx e λy f(y, c) dy [ ] x 0 ec x Ax y e λy dy dx = 1 mj=1 λ j nk=1 (A λ c) k with : A λ c 0; λ 0 with : R(A λ c) > 0; R(λ) > 0 10

11 Fenchel-duality: Laplace-duality f(b, c) = min{λ b f (λ, c)} f(b, c) = 1 λ Γ eλ b F (λ, c) dλ = min{b λ A (2iπ) m e λ b λ c; λ 0} = λ mj=1 λ nk=1 j (A dλ λ c) k Γ (2iπ) m where Γ C m Γ = u1 +i um and +i u 1 i u m i with u R m ; A u > c; u > 0. Polytope volume: > Brion, Brion and Vergnes, Barvinok and Pommersheim, Lasserre and Zeron,... Multidimensional integration over a simplex, ellipsoid > Lasserre and Zeron. 11

12 Remark: Logarithmic Barrier function of min{λ b A λ c) 0; λ 0} λ ϕ(r, λ) := rλ b n k=1 ln (A λ c) k m j=1 ln(λ j ). f(b, rc) = 1 r n f(rb, c) = Hence, lim r e min λ ϕ(r,λ) = = 1 r n (2iπ) m 1 r n (2iπ) m Γ e rλ b mj=1 λ j nk=1 (A λ c) k dλ Γ eϕ(r,λ) dλ lim min e ϕ(r,λ) 1 = lim r λ r r n (2iπ) m Γ eϕ(r,λ) dλ 12

13 Fenchel-duality: Laplace-duality f(b, c) = min{λ b f (λ, c)} f(b, c) = 1 λ = min λ {b λ A λ c; λ 0} = (2iπ) m (2iπ) m e λ b mj=1 Γ λ nk=1 j (A dλ λ c) k Γ = u1 +i Γ eλ b F (λ, c) dλ um where +i u 1 i u m i with u R m ; A u > c; u > 0. Simplex Algorithm Vertices of the polyhedron {Ax b; x 0} Cauchy s Residue Technique Poles of the function 1 mj=1 λ nk=1 j (A λ c) k THE SAME! 13

14 Laplace-Duality algorithm: Cauchy s Residue Technique f(b, c) = 1 (2iπ) m Γ e λ b mj=1 λ j nk=1 (A λ c) k dλ. Integration w.r.t. λ 1, λ 2,..., λ m (one variable at a time) by repeated application of Cauchy s residue theorem. Yields a tree of depth m whose level k is integration w.r.t. λ k on the integration path {u k i, u k + i } while the other variables λ k+1, λ k+2,..., λ m are fixed, on their respective integration path {u j i, u j + i }. Application of Cauchy s Residue Theorem yields a rational fraction in the variables λ k+1,..., λ m, and so on, until level m 1 where one obtains a rational fraction in the single variable λ m. 14

15 Brion and Vergne s continuous formula Terminology of LP in standard form: Let A σ := [A σ1... A σm ] be a basis of max{c x Ax = b; x 0}, with x(σ) the corresponding vertex, π σ := c σa 1 σ the associated dual variable and the reduced cost vector c k π σ A k, k σ. Then : f(b, c) = x(σ): vertex of Ω(b) from which it easily follows that e c x(σ) det(a σ ) k σ( c k + π σ A k ) log [ lim r f(b, rc) r] 1/r = max vertex of Ω(b) c x(σ). 15

16 Let Ω(y) := {x R n Ax = y; x 0}. Continuous Laplace-duality Discrete Z-duality f(y, c) := F (λ, c) := Ω(y) ec x dµ f d (y, c) := R m + e λy f(y, c) dy F d (z, c) := e c x x Ω(y) Z n y Z m z y f d (y, c) = 1 nk=1 (A λ c) k = n k=1 1 1 e c kz A 1k 1 z A mk m with R((A λ) k ) > c k k with z A 1k 1 z A mk m > e c k k 16

17 Continuous Laplace-duality Discrete Z-duality I. F (λ, c) f(b, c) I. F d (z, c) f d (b, c) by Inverse Laplace Transform and Cauchy s Residue Th. II. Data appear as by Inverse Z-Transform and Cauchy s Residue Th. II. Data appear as COEFFICIENTS EXPONENTS! of dual variables λ in F (λ, c) of dual variables z in F d (z, c). III. The poles of F (λ, c) are REAL III. The poles of F d (z, c) are COMPLEX and much more numerous! 17

18

19 Let σ := [A σ1 A σm ] be a feasible basis of {Ax = b, x 0}, with µ(σ) := det(a σ ), and dual variable π σ A σ = c σ. Continuous Laplace-duality basis σ poles : A σ λ = c σ a single real pole λ in R m Discrete Z-duality basis σ poles : z A σ = e c σ µ(σ) complex poles z = e λ in C m λ = π σ λ = π σ + i 2π v µ(σ) ; v V σ Z m V σ = {v Z m v A σ = 0 mod µ(σ)} 18

20 So both f d (b, c) and f(b, c) are computed with the same Cauchy s residue technique but the resulting contribution of each vertex for f d (b, c) is more complicated to evaluate because of the additional complex zeros. 19

21 Brion and Vergne s discrete formula Let σ := [A σ1 A σm ] be a feasible basis of {Ax = b, x 0}, with µ(σ) := det(a σ ), and µ(σ) dual variables z = e λ e 2iπv/µ(σ) C m with λ = π σ + i 2π v µ(σ) ; v V σ := {v Z m v A σ = 0 mod µ(σ)}. Re-interpreted with these data, Brion and Vergne s original discrete formula reads (Lasserre) f d (b, c) = e c x(σ) x(σ): vertex of Ω(b) 1 µ(σ) v V σ e2iπv b/µ(σ) k σ(1 e (2iπv A k /µ(σ)) e (c k π σ A k ) 20

22 Back to optimization : f d (b, c) = max{c x Ax = b; x N n }. Theorem max σ e c x(σ) lim r (Lasserre).Assume that v V σ k σ e2iπv b/µ(σ) (1 e (2iπv A k /µ(σ)) e r(c k π σ A k ) ) } is attained at a unique basis σ. Then : f d (b, c) = c x(σ ) + (c k π σ Ak )x k = k σ σ is an optimal basis of the linear program, and j c j x j. 1/r x(σ ) (resp. x ) is an optimal solution of the linear (resp. integer) program. 21

23 In this case : f d (b, c) = c x(σ ) + max A σ u k σ (c k π σ Ak )x k + k σ A k x k = b u Z m ; x k N k σ f d (b, c) = c x(σ ) + opt. value of GOMORY relaxation! 22

24 As a consequence, Corollary For t N sufficiently large, the gap f(tb, c) f d (tb, c) between the continuous and discrete optimal values is a (constant) periodic function with period µ(σ ) = det(a σ ), where σ is the optimal basis of the continous LP. The periodicity is due to the complex poles z = e λ with λ = π σ + i2π v µ(σ ), v V σ Zm ; v A σ = 0 mod µ(σ ). 23

25 KNAPSACK REVISITED Let a N n+1, 0 c R n+1, b N, and s := i a i. f d (b, c) = max x N n{c x a x = b}; f d (b, c) = x N n {e c x a x = b}. F d (z, c) = z s nj=1 (z a j e c j) ; f d (b, c) = 1 2iπ with γ R + and ln γ > max n i=1 c i/a i. z =γ zb 1 F d (z, c) dz The poles {z jk } of F satisfy : ln z jk = c j /a j + 2iπk/a j ; k = 1,..., a j ; j = 1,..., n 24

26 Poles of F(z) Integration path z = r 0 z = c_1/a_1 z = c_2/a_2 z =c_3/a_3 25 z = c_4/a_4

27 When the ratios {c k /a k } are close to each other, the integration of z b 1 F d (z) on the circle z = γ becomes more difficult because nearly all the poles of F d (z, c) contribute. Similarly, it is also known that the corresponding knapsack problems are difficult to solve... 26

28 Let r N, c 1 /a 1 < c 2 /a 2 < < c n /a n and c j N for all j. F d (z, rc) z = z s 1 nj=1 (z a j e rc j) = n j=1 Q j (z) (z a j e rc j), for some polynomials z Q j (z) := a j 1 k=0 Q jk z k = a j 1 k=0 [ Mjk (e r ) R jk (e r ) for some polynomials {M jk, R jk } of the variable y := e r. ] z k, In principle they can be obtained by symbolic calculation. Therefore, with b = s j mod a j, for all j = 1,..., n, f(b, rc) = n j=1 Q j(aj s j 1) y (b s j)c j a j = n j=1 M j(aj s j 1) (y) (b s j )c j R j(aj s j 1) (y) y a j 27

29 Hence e f d(b,c) = lim f r d (b, rc) 1/r = lim r n j=1 Thus, for sufficiently large b, e f d(b,c) = max j=1,...,n lim r M j(aj s j 1) (er (b s ) j )c j R j(aj s j 1) (er ) er a j M j(a j s j 1) (er ) R j(aj s j 1) (er ) er 1/r (b s j )c j a j 1/r f d (b, c) = c 1b a 1 c 1s 1 a 1 + deg M 1(a1 s 1 1) (y) deg R 1(a 1 s 1 1) (y) 28

30 Example : Consider the problem f(b, c) := max x N 2{3x 1 + 5x 2 2x 1 + 3x 2 b}. Let y = e r. We obtain F (z) z = z 5 (z 1)(z 2 y 3 )(z 3 y 5 ) = (y9 + y 7 + y 6 ) + z (y 7 + y 6 + y 4 ) + z 2 (y 6 + y 4 + y 2 ) (y 5 1)(y 1)(z 3 y 5 ) (y5 + y 3 ) + z (y 3 + y 2 ) (y 3 1)(y 1)(z 2 y 3 ) + 1 (y 5 1)(y 3 1)(z 1) Hence, f d (b, c) = f(b, c) if b = 0 mod 3 and for large b, f d (b, c) = f d (b, c) 5/3 + 1 whenever b = 1 mod 3 f d (b, c) = f d (b, c) 10/3 + 3 whenever b = 2 mod 3 29

31 A Discrete Farkas Lemma Let A N m n, b N m and consider the problem deciding whether or not Ax = b has a solution x N n. Theorem: (i) Ax = b has a solution x N n if and only if the polynomial b z b 1 in R[z 1,..., z m ] can be written z b 1 1 z b m 1 = n j=1 Q j (z)(z A j 1) = n j=1 Q j (z)(z A 1j 1 z A mj m 1) for some polynomials Q j (z), all with nonnegative coefficients. (ii) The degree of the Q j s is bounded by b := m j=1 b j max k m j=1 A jk. 30

32 A single LP to solve with n ( b ) ( +m b variables, b ) +m m constraints and a (sparse) matrix of coefficients in {0, ±1} One also retrieves the classical Farkas Lemma in R n, that is, {x R n Ax = b; x 0} = A u 0 b u 0. Indeed, if Ax = b has a solution x N n, then with u = ln z, Therefore, e b u 1 = n j=1 Q j (e u 1,..., e um )(e (A u) j 1). A u 0 e (A u) j 1 0 e b u 1 0 b u 0, and one retrieves (*), i.e., Ax = b has a solution x R n. 31

33 The general case A Z m n, b Z m. Let Ω := {x R n Ax = b; x 0} be a polytope. Let α N n be such that for every column A j of A, A kj +α j 0 k = 1,..., m; let N β ρ(α) := max{α x x Ω}. Theorem: (i) Ax = b has a solution x N n if and only if the polynomial z z b (zy) β 1 in R[z 1,..., z m ] can be written z b (zy) β 1 = Q 0 (z, y)(zy 1) + n j=1 Q j (z, y)(z A j(zy) α j 1), for some polynomials Q j (z), all with nonnegative coefficients. (ii) The degree of the Q j s is bounded by b := (m+1)β + m j=1 32 b j.

34 Back to standard Farkas lemma {x R n Ax = b, x 0} = [ A λ 0 ] b λ 0. But, equivalently {x R n Ax = b, x 0} = if and only if the polynomial λ b λ can be written b λ = n j=1 Q j (λ)(a λ) j, for some polynomials {Q j } R[λ 1,..., λ m ], all with nonnegative coefficients. In this case, each Q j is necessarily a constant, that is, Q j Q j (0) = x j 0, and Ax = b! 33

35 P = {x R n Ax = b, x 0} P Z n x P x = Q(0,..., 0) with x integer hull (P) x = Q(1,..., 1) with Q R[λ 1,..., λ m ] Q R[e λ 1,..., e λ m] b λ = Q, A λ e b λ 1 = Q, e A λ 1 n Q 0 Q 0 Comparing continuous and discrete Farkas lemma

36 An equivalent Linear program Let 0 q = {q jα } R ns be the coefficients of the Q j s in z b 1 = n j=1 Q j (z)(z A j 1) They are solutions of a linear system M q = r, q 0 for some matrix M and vector r, both with 0, ±1 coefficients. ** M and r are easily obtained from A, b with no computation Write q = (q 1, q 2,..., q n ) with each q j ĉ jα := c j for all α = {q jα } R s, and let 35

37 Theorem : Let A N m n, b N m, c R n. (i) The integer program P max{c x Ax = b, x N n } has same value as the linear program Q max{ n j=1 ĉ j q j M q = r; q 0}. (ii) Let q be an optimal vertex, and let x j := α q jα j = 1,..., n. Then x N n and x is an optimal solution of P. 36

38 The link with superadditive functions The LP-dual Q of the linear program Q reads Q min π {π r M π ĉ}. More precisely, with D := n j=1 {0, 1,..., b j } N m, Q min π π(b) π(0) s.t. π(α + A j ) π(α) c j, α D, j = 1,..., n Let Π := {π : N m R {+ } π(x) = if and only if x D}. For every π Π, let f π : N m R {+ } be the function f π (x) := inf α D π(α + x) π(α), x Nm 37

39 For every π Π, the function f π is superadditive and f π (0) = 0. The LP dual Q reads Q min π Π f π(b) s.t. f π (A j ) c j, j = 1,..., n. Thus, Q is a simplified and explicit form of the abstract dual of Jeroslow, Wolsey, stated in terms of superadditive functions. In the abstract dual one may restrict to the subclass of superadditive functions derived from the representation z b 1 = n j=1 Q j (z A j 1). 38

40 And, with P = {x R n + Ax = b} the integer hull co(p Zn ) reads co(p Z n ) = {x R n n j=1 f π (A j )x j f π (b)}, for finitely many π, generators of the convex cone (π, λ) : π(α + A j ) π(α) + λ j 0, α + A j D, j = 1,... n 39

41 CONCLUSION Generating functions permit to exhibit a natural duality for integer programming, an IP-analogue of LP duality. This duality also shows which kind of superadditive functions are useful in the abstract dual of Jeroslow, Wolsey. This might help providing efficient Gomory cuts in MIP solvers like CPLEX, or XPRESS-MP. 40

42 Another dual problem Let A Z m n, b Z m, c R n. Let y f(y, c) = max{c x Ax = y; x 0}. The Fenchel transform of the convex function f(., c) is the convex function λ ( f) (λ, c) = sup y R m λ y + f(y, c). The dual problem of the linear program is obtained from Fenchel duality as f(b, c) = inf λ R mb λ + ( f) ( λ, c) = inf λ R m b λ + sup x R m(c A λ) x = min {b λ A λ c} Equivalently e f(b,c) = inf λ R m sup e(b Ax) λ e c x x R m 41

43 Define Hence, ρ := inf z C m f d (z, c) = R z b sup R ( z b Ax ) e c x = inf f x N n z C m d (z, c). n j=1 (z A je c j) x j < if z A j e c j j (that is, A ln z c). Next, (writing z C as e λ e iθ ) ρ inf sup R ( z b Ax ) e c x = inf sup z R m x R n λ R m e(b Ax) λ e c x = e f(b,c) x R n Finally, with z C m arbitrary fixed sup R ( z b Ax ) e c x e c x = e f d(b,c) x N n Hence f d (b, c) ln ρ f(b, c). 42

44 Let σ be an optimal basis of the linear program. Under uniqueness of the max σ in Brion and Vergne s formula, and an additional technical condition e f d(b,c) = ρ = max x N n R ( ẑ b Ax e c x ) = f d (ẑ, c) where ẑ A j = γ e c j j σ for some real γ > 1. ẑ is an optimal solution of the dual problem inf f z C m d (z, c) 43

A non-standart approach to Duality

A non-standart approach to Duality Department of Industrial and Systems Engineering Lehigh University COR@L Seminar Series, 2005 10/27/2005 Outline 1 Duality Integration & Counting Duality for Integration Brion and Vergne s continuous formula

More information

DUALITY AND A FARKAS LEMMA FOR INTEGER PROGRAMS

DUALITY AND A FARKAS LEMMA FOR INTEGER PROGRAMS DUALITY AND A FARKAS LEMMA FOR INTEGER PROGRAMS JEAN B. LASSERRE Abstract. We consider the integer program max{c x Ax = b, x N n }. A formal parallel between linear programming and continuous integration

More information

Chapter 2 Duality and a Farkas lemma for integer programs

Chapter 2 Duality and a Farkas lemma for integer programs Chapter 2 Duality and a Farkas lemma for integer programs Jean B. Lasserre Abstract We consider the integer program max{c x Ax = b, x N n }.A formal parallel between linear programming and continuous integration,

More information

Integer programming, Barvinok s counting algorithm and Gomory relaxations

Integer programming, Barvinok s counting algorithm and Gomory relaxations Integer programming, Barvinok s counting algorithm and Gomory relaxations Jean B. Lasserre LAAS-CNRS, Toulouse, France Abstract We propose an algorithm based on Barvinok s counting algorithm for P max{c

More information

Integer Programming Duality

Integer Programming Duality Integer Programming Duality M. Guzelsoy T. K. Ralphs July, 2010 1 Introduction This article describes what is known about duality for integer programs. It is perhaps surprising that many of the results

More information

Springer Series in Operations Research and Financial Engineering. Series Editors: Thomas V. Mikosch Sidney I. Resnick Stephen M.

Springer Series in Operations Research and Financial Engineering. Series Editors: Thomas V. Mikosch Sidney I. Resnick Stephen M. Springer Series in Operations Research and Financial Engineering Series Editors: Thomas V. Mikosch Sidney I. Resnick Stephen M. Robinson Springer Series in Operations Research and Financial Engineering

More information

Duality for Mixed-Integer Linear Programs

Duality for Mixed-Integer Linear Programs Duality for Mixed-Integer Linear Programs M. Guzelsoy T. K. Ralphs Original May, 006 Revised April, 007 Abstract The theory of duality for linear programs is well-developed and has been successful in advancing

More information

Optimization WS 13/14:, by Y. Goldstein/K. Reinert, 9. Dezember 2013, 16: Linear programming. Optimization Problems

Optimization WS 13/14:, by Y. Goldstein/K. Reinert, 9. Dezember 2013, 16: Linear programming. Optimization Problems Optimization WS 13/14:, by Y. Goldstein/K. Reinert, 9. Dezember 2013, 16:38 2001 Linear programming Optimization Problems General optimization problem max{z(x) f j (x) 0,x D} or min{z(x) f j (x) 0,x D}

More information

Considering our result for the sum and product of analytic functions, this means that for (a 0, a 1,..., a N ) C N+1, the polynomial.

Considering our result for the sum and product of analytic functions, this means that for (a 0, a 1,..., a N ) C N+1, the polynomial. Lecture 3 Usual complex functions MATH-GA 245.00 Complex Variables Polynomials. Construction f : z z is analytic on all of C since its real and imaginary parts satisfy the Cauchy-Riemann relations and

More information

MIP reformulations of some chance-constrained mathematical programs

MIP reformulations of some chance-constrained mathematical programs MIP reformulations of some chance-constrained mathematical programs Ricardo Fukasawa Department of Combinatorics & Optimization University of Waterloo December 4th, 2012 FIELDS Industrial Optimization

More information

Cutting Planes in SCIP

Cutting Planes in SCIP Cutting Planes in SCIP Kati Wolter Zuse-Institute Berlin Department Optimization Berlin, 6th June 2007 Outline 1 Cutting Planes in SCIP 2 Cutting Planes for the 0-1 Knapsack Problem 2.1 Cover Cuts 2.2

More information

Integer Programming Methods LNMB

Integer Programming Methods LNMB Integer Programming Methods LNMB 2017 2018 Dion Gijswijt homepage.tudelft.nl/64a8q/intpm/ Dion Gijswijt Intro IntPM 2017-2018 1 / 24 Organisation Webpage: homepage.tudelft.nl/64a8q/intpm/ Book: Integer

More information

The master equality polyhedron with multiple rows

The master equality polyhedron with multiple rows The master equality polyhedron with multiple rows Sanjeeb Dash IBM Research sanjeebd@us.ibm.com Ricardo Fukasawa University of Waterloo rfukasaw@math.uwaterloo.ca September 16, 2010 Oktay Günlük IBM Research

More information

Lecture 9: Dantzig-Wolfe Decomposition

Lecture 9: Dantzig-Wolfe Decomposition Lecture 9: Dantzig-Wolfe Decomposition (3 units) Outline Dantzig-Wolfe decomposition Column generation algorithm Relation to Lagrangian dual Branch-and-price method Generated assignment problem and multi-commodity

More information

March 2002, December Introduction. We investigate the facial structure of the convex hull of the mixed integer knapsack set

March 2002, December Introduction. We investigate the facial structure of the convex hull of the mixed integer knapsack set ON THE FACETS OF THE MIXED INTEGER KNAPSACK POLYHEDRON ALPER ATAMTÜRK Abstract. We study the mixed integer knapsack polyhedron, that is, the convex hull of the mixed integer set defined by an arbitrary

More information

3. Linear Programming and Polyhedral Combinatorics

3. Linear Programming and Polyhedral Combinatorics Massachusetts Institute of Technology 18.433: Combinatorial Optimization Michel X. Goemans February 28th, 2013 3. Linear Programming and Polyhedral Combinatorics Summary of what was seen in the introductory

More information

1. Partial Fraction Expansion All the polynomials in this note are assumed to be complex polynomials.

1. Partial Fraction Expansion All the polynomials in this note are assumed to be complex polynomials. Partial Fraction Expansion All the polynomials in this note are assumed to be complex polynomials A rational function / is a quotient of two polynomials P, Q with 0 By Fundamental Theorem of algebra, =

More information

Lecture 8: Column Generation

Lecture 8: Column Generation Lecture 8: Column Generation (3 units) Outline Cutting stock problem Classical IP formulation Set covering formulation Column generation A dual perspective 1 / 24 Cutting stock problem 2 / 24 Problem description

More information

Combinatorial Optimization

Combinatorial Optimization Combinatorial Optimization Lecture notes, WS 2010/11, TU Munich Prof. Dr. Raymond Hemmecke Version of February 9, 2011 Contents 1 The knapsack problem 1 1.1 Complete enumeration..................................

More information

Asteroide Santana, Santanu S. Dey. December 4, School of Industrial and Systems Engineering, Georgia Institute of Technology

Asteroide Santana, Santanu S. Dey. December 4, School of Industrial and Systems Engineering, Georgia Institute of Technology for Some for Asteroide Santana, Santanu S. Dey School of Industrial Systems Engineering, Georgia Institute of Technology December 4, 2016 1 / 38 1 1.1 Conic integer programs for Conic integer programs

More information

Cutting Plane Methods I

Cutting Plane Methods I 6.859/15.083 Integer Programming and Combinatorial Optimization Fall 2009 Cutting Planes Consider max{wx : Ax b, x integer}. Cutting Plane Methods I Establishing the optimality of a solution is equivalent

More information

1. If 1, ω, ω 2, -----, ω 9 are the 10 th roots of unity, then (1 + ω) (1 + ω 2 ) (1 + ω 9 ) is A) 1 B) 1 C) 10 D) 0

1. If 1, ω, ω 2, -----, ω 9 are the 10 th roots of unity, then (1 + ω) (1 + ω 2 ) (1 + ω 9 ) is A) 1 B) 1 C) 10 D) 0 4 INUTES. If, ω, ω, -----, ω 9 are the th roots of unity, then ( + ω) ( + ω ) ----- ( + ω 9 ) is B) D) 5. i If - i = a + ib, then a =, b = B) a =, b = a =, b = D) a =, b= 3. Find the integral values for

More information

Lifting for conic mixed-integer programming

Lifting for conic mixed-integer programming Math. Program., Ser. A DOI 1.17/s117-9-282-9 FULL LENGTH PAPER Lifting for conic mixed-integer programming Alper Atamtürk Vishnu Narayanan Received: 13 March 28 / Accepted: 28 January 29 The Author(s)

More information

Computing the continuous discretely: The magic quest for a volume

Computing the continuous discretely: The magic quest for a volume Computing the continuous discretely: The magic quest for a volume Matthias Beck San Francisco State University math.sfsu.edu/beck Joint work with... Dennis Pixton (Birkhoff volume) Ricardo Diaz and Sinai

More information

Strong Dual for Conic Mixed-Integer Programs

Strong Dual for Conic Mixed-Integer Programs Strong Dual for Conic Mixed-Integer Programs Diego A. Morán R. Santanu S. Dey Juan Pablo Vielma July 14, 011 Abstract Mixed-integer conic programming is a generalization of mixed-integer linear programming.

More information

Chapter 1. Preliminaries

Chapter 1. Preliminaries Introduction This dissertation is a reading of chapter 4 in part I of the book : Integer and Combinatorial Optimization by George L. Nemhauser & Laurence A. Wolsey. The chapter elaborates links between

More information

Mixed-integer nonlinear programming, Conic programming, Duality, Cutting

Mixed-integer nonlinear programming, Conic programming, Duality, Cutting A STRONG DUAL FOR CONIC MIXED-INTEGER PROGRAMS DIEGO A. MORÁN R., SANTANU S. DEY, AND JUAN PABLO VIELMA Abstract. Mixed-integer conic programming is a generalization of mixed-integer linear programming.

More information

In particular, if A is a square matrix and λ is one of its eigenvalues, then we can find a non-zero column vector X with

In particular, if A is a square matrix and λ is one of its eigenvalues, then we can find a non-zero column vector X with Appendix: Matrix Estimates and the Perron-Frobenius Theorem. This Appendix will first present some well known estimates. For any m n matrix A = [a ij ] over the real or complex numbers, it will be convenient

More information

AN ALTERNATIVE ALGORITHM FOR COUNTING LATTICE POINTS IN A CONVEX POLYTOPE. 1. Introduction. Consider the (not necessarily compact) polyhedron

AN ALTERNATIVE ALGORITHM FOR COUNTING LATTICE POINTS IN A CONVEX POLYTOPE. 1. Introduction. Consider the (not necessarily compact) polyhedron AN ALTERNATIVE ALGORITHM FOR COUNTING LATTICE POINTS IN A CONVEX POLYTOPE JEAN B. LASSERRE AND EDUARDO S. ZERON Abstract. We provide an alternative algorithm for counting lattice points in the convex polytope

More information

3.10 Lagrangian relaxation

3.10 Lagrangian relaxation 3.10 Lagrangian relaxation Consider a generic ILP problem min {c t x : Ax b, Dx d, x Z n } with integer coefficients. Suppose Dx d are the complicating constraints. Often the linear relaxation and the

More information

I.3. LMI DUALITY. Didier HENRION EECI Graduate School on Control Supélec - Spring 2010

I.3. LMI DUALITY. Didier HENRION EECI Graduate School on Control Supélec - Spring 2010 I.3. LMI DUALITY Didier HENRION henrion@laas.fr EECI Graduate School on Control Supélec - Spring 2010 Primal and dual For primal problem p = inf x g 0 (x) s.t. g i (x) 0 define Lagrangian L(x, z) = g 0

More information

The master equality polyhedron with multiple rows

The master equality polyhedron with multiple rows The master equality polyhedron with multiple rows Sanjeeb Dash Ricardo Fukasawa IBM Research February 17, 2009 Oktay Günlük Abstract The master equality polyhedron (MEP) is a canonical set that generalizes

More information

Small (Explicit) Extended Formulation for Knapsack Cover Inequalities from Monotone Circuits

Small (Explicit) Extended Formulation for Knapsack Cover Inequalities from Monotone Circuits Small (Explicit) Extended Formulation for Knapsack Cover Inequalities from Monotone Circuits Abbas Bazzi Samuel Fiorini Sangxia Huang Ola Svensson & Samuel Fiorini Tony Huynh Stefan Weltge Extended formulations

More information

Section Notes 9. IP: Cutting Planes. Applied Math 121. Week of April 12, 2010

Section Notes 9. IP: Cutting Planes. Applied Math 121. Week of April 12, 2010 Section Notes 9 IP: Cutting Planes Applied Math 121 Week of April 12, 2010 Goals for the week understand what a strong formulations is. be familiar with the cutting planes algorithm and the types of cuts

More information

3. Linear Programming and Polyhedral Combinatorics

3. Linear Programming and Polyhedral Combinatorics Massachusetts Institute of Technology 18.453: Combinatorial Optimization Michel X. Goemans April 5, 2017 3. Linear Programming and Polyhedral Combinatorics Summary of what was seen in the introductory

More information

Introduction to optimization

Introduction to optimization Introduction to optimization Geir Dahl CMA, Dept. of Mathematics and Dept. of Informatics University of Oslo 1 / 24 The plan 1. The basic concepts 2. Some useful tools (linear programming = linear optimization)

More information

Linear Programming. Chapter Introduction

Linear Programming. Chapter Introduction Chapter 3 Linear Programming Linear programs (LP) play an important role in the theory and practice of optimization problems. Many COPs can directly be formulated as LPs. Furthermore, LPs are invaluable

More information

MVE165/MMG631 Linear and integer optimization with applications Lecture 8 Discrete optimization: theory and algorithms

MVE165/MMG631 Linear and integer optimization with applications Lecture 8 Discrete optimization: theory and algorithms MVE165/MMG631 Linear and integer optimization with applications Lecture 8 Discrete optimization: theory and algorithms Ann-Brith Strömberg 2017 04 07 Lecture 8 Linear and integer optimization with applications

More information

On the knapsack closure of 0-1 Integer Linear Programs. Matteo Fischetti University of Padova, Italy

On the knapsack closure of 0-1 Integer Linear Programs. Matteo Fischetti University of Padova, Italy On the knapsack closure of 0-1 Integer Linear Programs Matteo Fischetti University of Padova, Italy matteo.fischetti@unipd.it Andrea Lodi University of Bologna, Italy alodi@deis.unibo.it Aussois, January

More information

Polyhedral Approach to Integer Linear Programming. Tepper School of Business Carnegie Mellon University, Pittsburgh

Polyhedral Approach to Integer Linear Programming. Tepper School of Business Carnegie Mellon University, Pittsburgh Polyhedral Approach to Integer Linear Programming Gérard Cornuéjols Tepper School of Business Carnegie Mellon University, Pittsburgh 1 / 30 Brief history First Algorithms Polynomial Algorithms Solving

More information

International Competition in Mathematics for Universtiy Students in Plovdiv, Bulgaria 1994

International Competition in Mathematics for Universtiy Students in Plovdiv, Bulgaria 1994 International Competition in Mathematics for Universtiy Students in Plovdiv, Bulgaria 1994 1 PROBLEMS AND SOLUTIONS First day July 29, 1994 Problem 1. 13 points a Let A be a n n, n 2, symmetric, invertible

More information

3.7 Cutting plane methods

3.7 Cutting plane methods 3.7 Cutting plane methods Generic ILP problem min{ c t x : x X = {x Z n + : Ax b} } with m n matrix A and n 1 vector b of rationals. According to Meyer s theorem: There exists an ideal formulation: conv(x

More information

BCOL RESEARCH REPORT 07.04

BCOL RESEARCH REPORT 07.04 BCOL RESEARCH REPORT 07.04 Industrial Engineering & Operations Research University of California, Berkeley, CA 94720-1777 LIFTING FOR CONIC MIXED-INTEGER PROGRAMMING ALPER ATAMTÜRK AND VISHNU NARAYANAN

More information

Lecture 8: Column Generation

Lecture 8: Column Generation Lecture 8: Column Generation (3 units) Outline Cutting stock problem Classical IP formulation Set covering formulation Column generation A dual perspective Vehicle routing problem 1 / 33 Cutting stock

More information

Some Relationships between Disjunctive Cuts and Cuts based on S-free Convex Sets

Some Relationships between Disjunctive Cuts and Cuts based on S-free Convex Sets Some Relationships between Disjunctive Cuts and Cuts based on S-free Convex Sets Sanjeeb Dash a Santanu S. Dey b Oktay Günlük a a Business Analytics and Mathematical Sciences Department, IBM T. J. Watson

More information

The partial-fractions method for counting solutions to integral linear systems

The partial-fractions method for counting solutions to integral linear systems The partial-fractions method for counting solutions to integral linear systems Matthias Beck, MSRI www.msri.org/people/members/matthias/ arxiv: math.co/0309332 Vector partition functions A an (m d)-integral

More information

Lagrangian Relaxation in MIP

Lagrangian Relaxation in MIP Lagrangian Relaxation in MIP Bernard Gendron May 28, 2016 Master Class on Decomposition, CPAIOR2016, Banff, Canada CIRRELT and Département d informatique et de recherche opérationnelle, Université de Montréal,

More information

Sequence independent lifting for 0 1 knapsack problems with disjoint cardinality constraints

Sequence independent lifting for 0 1 knapsack problems with disjoint cardinality constraints Sequence independent lifting for knapsack problems with disjoint cardinality constraints Bo Zeng and Jean-Philippe P. Richard School of Industrial Engineering, Purdue University, 35 N. Grant Street, West

More information

VALID INEQUALITIES FOR MIXED-INTEGER LINEAR PROGRAMMING PROBLEMS

VALID INEQUALITIES FOR MIXED-INTEGER LINEAR PROGRAMMING PROBLEMS VALID INEQUALITIES FOR MIXED-INTEGER LINEAR PROGRAMMING PROBLEMS BY EMRE YAMANGIL A dissertation submitted to the Graduate School New Brunswick Rutgers, The State University of New Jersey in partial fulfillment

More information

On John type ellipsoids

On John type ellipsoids On John type ellipsoids B. Klartag Tel Aviv University Abstract Given an arbitrary convex symmetric body K R n, we construct a natural and non-trivial continuous map u K which associates ellipsoids to

More information

A notion of Total Dual Integrality for Convex, Semidefinite and Extended Formulations

A notion of Total Dual Integrality for Convex, Semidefinite and Extended Formulations A notion of for Convex, Semidefinite and Extended Formulations Marcel de Carli Silva Levent Tunçel April 26, 2018 A vector in R n is integral if each of its components is an integer, A vector in R n is

More information

Integer Programming ISE 418. Lecture 12. Dr. Ted Ralphs

Integer Programming ISE 418. Lecture 12. Dr. Ted Ralphs Integer Programming ISE 418 Lecture 12 Dr. Ted Ralphs ISE 418 Lecture 12 1 Reading for This Lecture Nemhauser and Wolsey Sections II.2.1 Wolsey Chapter 9 ISE 418 Lecture 12 2 Generating Stronger Valid

More information

Integer Programming ISE 418. Lecture 8. Dr. Ted Ralphs

Integer Programming ISE 418. Lecture 8. Dr. Ted Ralphs Integer Programming ISE 418 Lecture 8 Dr. Ted Ralphs ISE 418 Lecture 8 1 Reading for This Lecture Wolsey Chapter 2 Nemhauser and Wolsey Sections II.3.1, II.3.6, II.4.1, II.4.2, II.5.4 Duality for Mixed-Integer

More information

Polynomiality of Linear Programming

Polynomiality of Linear Programming Chapter 10 Polynomiality of Linear Programming In the previous section, we presented the Simplex Method. This method turns out to be very efficient for solving linear programmes in practice. While it is

More information

Cutting Plane Methods II

Cutting Plane Methods II 6.859/5.083 Integer Programming and Combinatorial Optimization Fall 2009 Cutting Plane Methods II Gomory-Chvátal cuts Reminder P = {x R n : Ax b} with A Z m n, b Z m. For λ [0, ) m such that λ A Z n, (λ

More information

CO 250 Final Exam Guide

CO 250 Final Exam Guide Spring 2017 CO 250 Final Exam Guide TABLE OF CONTENTS richardwu.ca CO 250 Final Exam Guide Introduction to Optimization Kanstantsin Pashkovich Spring 2017 University of Waterloo Last Revision: March 4,

More information

Linear and Integer Optimization (V3C1/F4C1)

Linear and Integer Optimization (V3C1/F4C1) Linear and Integer Optimization (V3C1/F4C1) Lecture notes Ulrich Brenner Research Institute for Discrete Mathematics, University of Bonn Winter term 2016/2017 March 8, 2017 12:02 1 Preface Continuous updates

More information

A Review of Linear Programming

A Review of Linear Programming A Review of Linear Programming Instructor: Farid Alizadeh IEOR 4600y Spring 2001 February 14, 2001 1 Overview In this note we review the basic properties of linear programming including the primal simplex

More information

Introduction to Integer Linear Programming

Introduction to Integer Linear Programming Lecture 7/12/2006 p. 1/30 Introduction to Integer Linear Programming Leo Liberti, Ruslan Sadykov LIX, École Polytechnique liberti@lix.polytechnique.fr sadykov@lix.polytechnique.fr Lecture 7/12/2006 p.

More information

MINI-TUTORIAL ON SEMI-ALGEBRAIC PROOF SYSTEMS. Albert Atserias Universitat Politècnica de Catalunya Barcelona

MINI-TUTORIAL ON SEMI-ALGEBRAIC PROOF SYSTEMS. Albert Atserias Universitat Politècnica de Catalunya Barcelona MINI-TUTORIAL ON SEMI-ALGEBRAIC PROOF SYSTEMS Albert Atserias Universitat Politècnica de Catalunya Barcelona Part I CONVEX POLYTOPES Convex polytopes as linear inequalities Polytope: P = {x R n : Ax b}

More information

3 (Due ). Let A X consist of points (x, y) such that either x or y is a rational number. Is A measurable? What is its Lebesgue measure?

3 (Due ). Let A X consist of points (x, y) such that either x or y is a rational number. Is A measurable? What is its Lebesgue measure? MA 645-4A (Real Analysis), Dr. Chernov Homework assignment 1 (Due ). Show that the open disk x 2 + y 2 < 1 is a countable union of planar elementary sets. Show that the closed disk x 2 + y 2 1 is a countable

More information

Appendix PRELIMINARIES 1. THEOREMS OF ALTERNATIVES FOR SYSTEMS OF LINEAR CONSTRAINTS

Appendix PRELIMINARIES 1. THEOREMS OF ALTERNATIVES FOR SYSTEMS OF LINEAR CONSTRAINTS Appendix PRELIMINARIES 1. THEOREMS OF ALTERNATIVES FOR SYSTEMS OF LINEAR CONSTRAINTS Here we consider systems of linear constraints, consisting of equations or inequalities or both. A feasible solution

More information

Cutting Plane Separators in SCIP

Cutting Plane Separators in SCIP Cutting Plane Separators in SCIP Kati Wolter Zuse Institute Berlin DFG Research Center MATHEON Mathematics for key technologies 1 / 36 General Cutting Plane Method MIP min{c T x : x X MIP }, X MIP := {x

More information

Linear Programming. Scheduling problems

Linear Programming. Scheduling problems Linear Programming Scheduling problems Linear programming (LP) ( )., 1, for 0 min 1 1 1 1 1 11 1 1 n i x b x a x a b x a x a x c x c x z i m n mn m n n n n! = + + + + + + = Extreme points x ={x 1,,x n

More information

Notes on Complex Analysis

Notes on Complex Analysis Michael Papadimitrakis Notes on Complex Analysis Department of Mathematics University of Crete Contents The complex plane.. The complex plane...................................2 Argument and polar representation.........................

More information

Multicommodity Flows and Column Generation

Multicommodity Flows and Column Generation Lecture Notes Multicommodity Flows and Column Generation Marc Pfetsch Zuse Institute Berlin pfetsch@zib.de last change: 2/8/2006 Technische Universität Berlin Fakultät II, Institut für Mathematik WS 2006/07

More information

CSCI 1951-G Optimization Methods in Finance Part 01: Linear Programming

CSCI 1951-G Optimization Methods in Finance Part 01: Linear Programming CSCI 1951-G Optimization Methods in Finance Part 01: Linear Programming January 26, 2018 1 / 38 Liability/asset cash-flow matching problem Recall the formulation of the problem: max w c 1 + p 1 e 1 = 150

More information

Math Homework 2

Math Homework 2 Math 73 Homework Due: September 8, 6 Suppose that f is holomorphic in a region Ω, ie an open connected set Prove that in any of the following cases (a) R(f) is constant; (b) I(f) is constant; (c) f is

More information

Convex Optimization & Parsimony of L p-balls representation

Convex Optimization & Parsimony of L p-balls representation Convex Optimization & Parsimony of L p -balls representation LAAS-CNRS and Institute of Mathematics, Toulouse, France IMA, January 2016 Motivation Unit balls associated with nonnegative homogeneous polynomials

More information

Computational Integer Programming Universidad de los Andes. Lecture 1. Dr. Ted Ralphs

Computational Integer Programming Universidad de los Andes. Lecture 1. Dr. Ted Ralphs Computational Integer Programming Universidad de los Andes Lecture 1 Dr. Ted Ralphs MIP Lecture 1 1 Quick Introduction Bio Course web site Course structure http://coral.ie.lehigh.edu/ ted/teaching/mip

More information

n-step mingling inequalities: new facets for the mixed-integer knapsack set

n-step mingling inequalities: new facets for the mixed-integer knapsack set Math. Program., Ser. A (2012) 132:79 98 DOI 10.1007/s10107-010-0382-6 FULL LENGTH PAPER n-step mingling inequalities: new facets for the mixed-integer knapsack set Alper Atamtürk Kiavash Kianfar Received:

More information

(1) Consider the space S consisting of all continuous real-valued functions on the closed interval [0, 1]. For f, g S, define

(1) Consider the space S consisting of all continuous real-valued functions on the closed interval [0, 1]. For f, g S, define Homework, Real Analysis I, Fall, 2010. (1) Consider the space S consisting of all continuous real-valued functions on the closed interval [0, 1]. For f, g S, define ρ(f, g) = 1 0 f(x) g(x) dx. Show that

More information

Submodular Inequalities for the Path Structures of the Capacitated Fixed-Charge Network Flow Problems. Birce Tezel

Submodular Inequalities for the Path Structures of the Capacitated Fixed-Charge Network Flow Problems. Birce Tezel Submodular Inequalities for the Path Structures of the Capacitated Fixed-Charge Network Flow Problems By Birce Tezel A dissertation submitted in partial satisfaction of the requirements for the degree

More information

Introduction to Semidefinite Programming I: Basic properties a

Introduction to Semidefinite Programming I: Basic properties a Introduction to Semidefinite Programming I: Basic properties and variations on the Goemans-Williamson approximation algorithm for max-cut MFO seminar on Semidefinite Programming May 30, 2010 Semidefinite

More information

Farkas Lemma. Rudi Pendavingh. Optimization in R n, lecture 2. Eindhoven Technical University. Rudi Pendavingh (TUE) Farkas Lemma ORN2 1 / 15

Farkas Lemma. Rudi Pendavingh. Optimization in R n, lecture 2. Eindhoven Technical University. Rudi Pendavingh (TUE) Farkas Lemma ORN2 1 / 15 Farkas Lemma Rudi Pendavingh Eindhoven Technical University Optimization in R n, lecture 2 Rudi Pendavingh (TUE) Farkas Lemma ORN2 1 / 15 Today s Lecture Theorem (Farkas Lemma, 1894) Let A be an m n matrix,

More information

The Value function of a Mixed-Integer Linear Program with a Single Constraint

The Value function of a Mixed-Integer Linear Program with a Single Constraint The Value Function of a Mixed Integer Linear Programs with a Single Constraint MENAL GUZELSOY TED RALPHS ISE Department COR@L Lab Lehigh University tkralphs@lehigh.edu University of Wisconsin February

More information

Optimization Methods in Management Science

Optimization Methods in Management Science Optimization Methods in Management Science MIT 15.05 Recitation 8 TAs: Giacomo Nannicini, Ebrahim Nasrabadi At the end of this recitation, students should be able to: 1. Derive Gomory cut from fractional

More information

On mixed-integer sets with two integer variables

On mixed-integer sets with two integer variables On mixed-integer sets with two integer variables Sanjeeb Dash IBM Research sanjeebd@us.ibm.com Santanu S. Dey Georgia Inst. Tech. santanu.dey@isye.gatech.edu September 8, 2010 Oktay Günlük IBM Research

More information

Lecture Note 5: Semidefinite Programming for Stability Analysis

Lecture Note 5: Semidefinite Programming for Stability Analysis ECE7850: Hybrid Systems:Theory and Applications Lecture Note 5: Semidefinite Programming for Stability Analysis Wei Zhang Assistant Professor Department of Electrical and Computer Engineering Ohio State

More information

Lectures 6, 7 and part of 8

Lectures 6, 7 and part of 8 Lectures 6, 7 and part of 8 Uriel Feige April 26, May 3, May 10, 2015 1 Linear programming duality 1.1 The diet problem revisited Recall the diet problem from Lecture 1. There are n foods, m nutrients,

More information

7. Lecture notes on the ellipsoid algorithm

7. Lecture notes on the ellipsoid algorithm Massachusetts Institute of Technology Michel X. Goemans 18.433: Combinatorial Optimization 7. Lecture notes on the ellipsoid algorithm The simplex algorithm was the first algorithm proposed for linear

More information

Column basis reduction and decomposable knapsack problems

Column basis reduction and decomposable knapsack problems Column basis reduction and decomposable knapsack problems Bala Krishnamoorthy Gábor Pataki Abstract We propose a very simple preconditioning method for integer programming feasibility problems: replacing

More information

Structure of Valid Inequalities for Mixed Integer Conic Programs

Structure of Valid Inequalities for Mixed Integer Conic Programs Structure of Valid Inequalities for Mixed Integer Conic Programs Fatma Kılınç-Karzan Tepper School of Business Carnegie Mellon University 18 th Combinatorial Optimization Workshop Aussois, France January

More information

MATH 131A: REAL ANALYSIS (BIG IDEAS)

MATH 131A: REAL ANALYSIS (BIG IDEAS) MATH 131A: REAL ANALYSIS (BIG IDEAS) Theorem 1 (The Triangle Inequality). For all x, y R we have x + y x + y. Proposition 2 (The Archimedean property). For each x R there exists an n N such that n > x.

More information

A geometric perspective on lifting

A geometric perspective on lifting A geometric perspective on lifting Michele Conforti Università di Padova, conforti@math.unipd.it Gérard Cornuéjols Carnegie Mellon University and Université d Aix-Marseille, gc0v@andrew.cmu.edu Giacomo

More information

On the facets of the mixed integer knapsack polyhedron

On the facets of the mixed integer knapsack polyhedron Math. Program., Ser. B 98: 145 175 (2003) Digital Object Identifier (DOI) 10.1007/s10107-003-0400-z Alper Atamtürk On the facets of the mixed integer knapsack polyhedron Received: April 8, 2002 / Accepted:

More information

4. Algebra and Duality

4. Algebra and Duality 4-1 Algebra and Duality P. Parrilo and S. Lall, CDC 2003 2003.12.07.01 4. Algebra and Duality Example: non-convex polynomial optimization Weak duality and duality gap The dual is not intrinsic The cone

More information

Sequential pairing of mixed integer inequalities

Sequential pairing of mixed integer inequalities Sequential pairing of mixed integer inequalities Yongpei Guan, Shabbir Ahmed, George L. Nemhauser School of Industrial & Systems Engineering, Georgia Institute of Technology, 765 Ferst Drive, Atlanta,

More information

Lecture 5. Theorems of Alternatives and Self-Dual Embedding

Lecture 5. Theorems of Alternatives and Self-Dual Embedding IE 8534 1 Lecture 5. Theorems of Alternatives and Self-Dual Embedding IE 8534 2 A system of linear equations may not have a solution. It is well known that either Ax = c has a solution, or A T y = 0, c

More information

Integer Programming: Cutting Planes

Integer Programming: Cutting Planes OptIntro 1 / 39 Integer Programming: Cutting Planes Eduardo Camponogara Department of Automation and Systems Engineering Federal University of Santa Catarina October 2016 OptIntro 2 / 39 Summary Introduction

More information

4th Preparation Sheet - Solutions

4th Preparation Sheet - Solutions Prof. Dr. Rainer Dahlhaus Probability Theory Summer term 017 4th Preparation Sheet - Solutions Remark: Throughout the exercise sheet we use the two equivalent definitions of separability of a metric space

More information

A BRANCH&BOUND ALGORITHM FOR SOLVING ONE-DIMENSIONAL CUTTING STOCK PROBLEMS EXACTLY

A BRANCH&BOUND ALGORITHM FOR SOLVING ONE-DIMENSIONAL CUTTING STOCK PROBLEMS EXACTLY APPLICATIONES MATHEMATICAE 23,2 (1995), pp. 151 167 G. SCHEITHAUER and J. TERNO (Dresden) A BRANCH&BOUND ALGORITHM FOR SOLVING ONE-DIMENSIONAL CUTTING STOCK PROBLEMS EXACTLY Abstract. Many numerical computations

More information

LP Duality: outline. Duality theory for Linear Programming. alternatives. optimization I Idea: polyhedra

LP Duality: outline. Duality theory for Linear Programming. alternatives. optimization I Idea: polyhedra LP Duality: outline I Motivation and definition of a dual LP I Weak duality I Separating hyperplane theorem and theorems of the alternatives I Strong duality and complementary slackness I Using duality

More information

Theory and Internet Protocols

Theory and Internet Protocols Game Lecture 2: Linear Programming and Zero Sum Nash Equilibrium Xiaotie Deng AIMS Lab Department of Computer Science Shanghai Jiaotong University September 26, 2016 1 2 3 4 Standard Form (P) Outline

More information

Structure in Mixed Integer Conic Optimization: From Minimal Inequalities to Conic Disjunctive Cuts

Structure in Mixed Integer Conic Optimization: From Minimal Inequalities to Conic Disjunctive Cuts Structure in Mixed Integer Conic Optimization: From Minimal Inequalities to Conic Disjunctive Cuts Fatma Kılınç-Karzan Tepper School of Business Carnegie Mellon University Joint work with Sercan Yıldız

More information

Deciding Emptiness of the Gomory-Chvátal Closure is NP-Complete, Even for a Rational Polyhedron Containing No Integer Point

Deciding Emptiness of the Gomory-Chvátal Closure is NP-Complete, Even for a Rational Polyhedron Containing No Integer Point Deciding Emptiness of the Gomory-Chvátal Closure is NP-Complete, Even for a Rational Polyhedron Containing No Integer Point Gérard Cornuéjols 1 and Yanjun Li 2 1 Tepper School of Business, Carnegie Mellon

More information

HW1 solutions. 1. α Ef(x) β, where Ef(x) is the expected value of f(x), i.e., Ef(x) = n. i=1 p if(a i ). (The function f : R R is given.

HW1 solutions. 1. α Ef(x) β, where Ef(x) is the expected value of f(x), i.e., Ef(x) = n. i=1 p if(a i ). (The function f : R R is given. HW1 solutions Exercise 1 (Some sets of probability distributions.) Let x be a real-valued random variable with Prob(x = a i ) = p i, i = 1,..., n, where a 1 < a 2 < < a n. Of course p R n lies in the standard

More information

Hilbert s 17th Problem to Semidefinite Programming & Convex Algebraic Geometry

Hilbert s 17th Problem to Semidefinite Programming & Convex Algebraic Geometry Hilbert s 17th Problem to Semidefinite Programming & Convex Algebraic Geometry Rekha R. Thomas University of Washington, Seattle References Monique Laurent, Sums of squares, moment matrices and optimization

More information

Semidefinite Programming

Semidefinite Programming Chapter 2 Semidefinite Programming 2.0.1 Semi-definite programming (SDP) Given C M n, A i M n, i = 1, 2,..., m, and b R m, the semi-definite programming problem is to find a matrix X M n for the optimization

More information

Dirichlet s Theorem. Calvin Lin Zhiwei. August 18, 2007

Dirichlet s Theorem. Calvin Lin Zhiwei. August 18, 2007 Dirichlet s Theorem Calvin Lin Zhiwei August 8, 2007 Abstract This paper provides a proof of Dirichlet s theorem, which states that when (m, a) =, there are infinitely many primes uch that p a (mod m).

More information