DOCSLIB.ORG

Duality, Part 1: Dual Bases and Dual Maps Notation

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Duality, Part 1: Dual Bases and Dual Maps Notation Duality, part 1: Dual Bases and Dual Maps Notation F denotes either R or C. V and W denote vector spaces over F. Define ': R3 ! R by '(x; y; z) = 4x − 5y + 2z. Then ' is a linear functional on R3. n n Fix (b1;:::; bn) 2 C . Define ': C ! C by '(z1;:::; zn) = b1z1 + ··· + bnzn: Then ' is a linear functional on Cn. Define ': P(R) ! R by '(p) = 3p00(5) + 7p(4). Then ' is a linear functional on P(R). R 1 Define ': P(R) ! R by '(p) = 0 p(x) dx. Then ' is a linear functional on P(R). Examples: Linear Functionals Definition: linear functional A linear functional on V is a linear map from V to F. In other words, a linear functional is an element of L(V; F). n n Fix (b1;:::; bn) 2 C . Define ': C ! C by '(z1;:::; zn) = b1z1 + ··· + bnzn: Then ' is a linear functional on Cn. Define ': P(R) ! R by '(p) = 3p00(5) + 7p(4). Then ' is a linear functional on P(R). R 1 Define ': P(R) ! R by '(p) = 0 p(x) dx. Then ' is a linear functional on P(R). Linear Functionals Definition: linear functional A linear functional on V is a linear map from V to F. In other words, a linear functional is an element of L(V; F). Examples: Define ': R3 ! R by '(x; y; z) = 4x − 5y + 2z. Then ' is a linear functional on R3. Define ': P(R) ! R by '(p) = 3p00(5) + 7p(4). Then ' is a linear functional on P(R). R 1 Define ': P(R) ! R by '(p) = 0 p(x) dx. Then ' is a linear functional on P(R). Linear Functionals Definition: linear functional A linear functional on V is a linear map from V to F. In other words, a linear functional is an element of L(V; F). Examples: Define ': R3 ! R by '(x; y; z) = 4x − 5y + 2z. Then ' is a linear functional on R3. n n Fix (b1;:::; bn) 2 C . Define ': C ! C by '(z1;:::; zn) = b1z1 + ··· + bnzn: Then ' is a linear functional on Cn. R 1 Define ': P(R) ! R by '(p) = 0 p(x) dx. Then ' is a linear functional on P(R). Linear Functionals Definition: linear functional A linear functional on V is a linear map from V to F. In other words, a linear functional is an element of L(V; F). Examples: Define ': R3 ! R by '(x; y; z) = 4x − 5y + 2z. Then ' is a linear functional on R3. n n Fix (b1;:::; bn) 2 C . Define ': C ! C by '(z1;:::; zn) = b1z1 + ··· + bnzn: Then ' is a linear functional on Cn. Define ': P(R) ! R by '(p) = 3p00(5) + 7p(4). Then ' is a linear functional on P(R). Linear Functionals Definition: linear functional A linear functional on V is a linear map from V to F. In other words, a linear functional is an element of L(V; F). Examples: Define ': R3 ! R by '(x; y; z) = 4x − 5y + 2z. Then ' is a linear functional on R3. n n Fix (b1;:::; bn) 2 C . Define ': C ! C by '(z1;:::; zn) = b1z1 + ··· + bnzn: Then ' is a linear functional on Cn. Define ': P(R) ! R by '(p) = 3p00(5) + 7p(4). Then ' is a linear functional on P(R). R 1 Define ': P(R) ! R by '(p) = 0 p(x) dx. Then ' is a linear functional on P(R). = (dim V)(dim F) = dim V dim V0 = dim V Suppose V is finite-dimensional. Then V0 is also finite-dimensional and dim V0 = dim V. Proof dim V0 = dim L(V; F) The Dual Space and Its Dimension Definition: dual space, V0 The dual space of V, denoted V0, is the vector space of all linear functionals on V. In other words, V0 = L(V; F). = (dim V)(dim F) = dim V Proof dim V0 = dim L(V; F) The Dual Space and Its Dimension Definition: dual space, V0 The dual space of V, denoted V0, is the vector space of all linear functionals on V. In other words, V0 = L(V; F). dim V0 = dim V Suppose V is finite-dimensional. Then V0 is also finite-dimensional and dim V0 = dim V. = (dim V)(dim F) = dim V The Dual Space and Its Dimension Definition: dual space, V0 The dual space of V, denoted V0, is the vector space of all linear functionals on V. In other words, V0 = L(V; F). dim V0 = dim V Suppose V is finite-dimensional. Then V0 is also finite-dimensional and dim V0 = dim V. Proof dim V0 = dim L(V; F) = dim V The Dual Space and Its Dimension Definition: dual space, V0 The dual space of V, denoted V0, is the vector space of all linear functionals on V. In other words, V0 = L(V; F). dim V0 = dim V Suppose V is finite-dimensional. Then V0 is also finite-dimensional and dim V0 = dim V. Proof dim V0 = dim L(V; F) = (dim V)(dim F) The Dual Space and Its Dimension Definition: dual space, V0 The dual space of V, denoted V0, is the vector space of all linear functionals on V. In other words, V0 = L(V; F). dim V0 = dim V Suppose V is finite-dimensional. Then V0 is also finite-dimensional and dim V0 = dim V. Proof dim V0 = dim L(V; F) = (dim V)(dim F) = dim V Example: Consider the standard basis e1;:::; en of Fn. For 1 ≤ j ≤ n, define ' by Dual basis is a basis of j the dual space 'j(x1;:::; xn) = xj for (x ;:::; x ) 2 Fn. Clearly ( Suppose dim V < 1. Then 1 n 1 if k = j; 'j(ek) = the dual basis of a basis of V 6= : 0 if k j is a basis of V0. Thus '1;:::;'n is the dual basis of the n standard basis e1;:::; en of F . The Dual Basis Definition: dual basis If v1;:::; vn is a basis of V, then the dual basis of v1;:::; vn is the list 0 '1;:::;'n of elements of V , where each 'j is the linear functional on such that V ( 1 if k = j; 'j(vk) = 0 if k 6= j: For 1 ≤ j ≤ n, define ' by Dual basis is a basis of j the dual space 'j(x1;:::; xn) = xj for (x ;:::; x ) 2 Fn. Clearly ( Suppose dim V < 1. Then 1 n 1 if k = j; 'j(ek) = the dual basis of a basis of V 6= : 0 if k j is a basis of V0. Thus '1;:::;'n is the dual basis of the n standard basis e1;:::; en of F . The Dual Basis Definition: dual basis If v1;:::; vn is a basis of V, then the dual basis of v1;:::; vn is the list 0 '1;:::;'n of elements of V , where each 'j is the linear functional on such that V ( 1 if k = j; 'j(vk) = 0 if k 6= j: Example: Consider the standard basis e1;:::; en of Fn. Dual basis is a basis of the dual space Clearly ( Suppose dim V < 1. Then 1 if k = j; 'j(ek) = the dual basis of a basis of V 6= : 0 if k j is a basis of V0. Thus '1;:::;'n is the dual basis of the n standard basis e1;:::; en of F . The Dual Basis Definition: dual basis If v1;:::; vn is a basis of V, then the dual basis of v1;:::; vn is the list 0 '1;:::;'n of elements of V , where each 'j is the linear functional on such that V ( 1 if k = j; 'j(vk) = 0 if k 6= j: Example: Consider the standard basis e1;:::; en n of F . For 1 ≤ j ≤ n, define 'j by 'j(x1;:::; xn) = xj n for (x1;:::; xn) 2 F . Dual basis is a basis of the dual space Suppose dim V < 1. Then the dual basis of a basis of V is a basis of V0. The Dual Basis Definition: dual basis If v1;:::; vn is a basis of V, then the dual basis of v1;:::; vn is the list 0 '1;:::;'n of elements of V , where each 'j is the linear functional on such that V ( 1 if k = j; 'j(vk) = 0 if k 6= j: Example: Consider the standard basis e1;:::; en n of F . For 1 ≤ j ≤ n, define 'j by 'j(x1;:::; xn) = xj for (x ;:::; x ) 2 Fn. Clearly ( 1 n 1 if k = j; 'j(ek) = 0 if k 6= j: Thus '1;:::;'n is the dual basis of the n standard basis e1;:::; en of F . The Dual Basis Definition: dual basis If v1;:::; vn is a basis of V, then the dual basis of v1;:::; vn is the list 0 '1;:::;'n of elements of V , where each 'j is the linear functional on such that V ( 1 if k = j; 'j(vk) = 0 if k 6= j: Example: Consider the standard basis e1;:::; en of Fn. For 1 ≤ j ≤ n, define ' by Dual basis is a basis of j the dual space 'j(x1;:::; xn) = xj for (x ;:::; x ) 2 Fn. Clearly ( Suppose dim V < 1. Then 1 n 1 if k = j; 'j(ek) = the dual basis of a basis of V 6= : 0 if k j is a basis of V0. Thus '1;:::;'n is the dual basis of the n standard basis e1;:::; en of F . Example: Define D: P(R) !P(R) by Dp = p0. Suppose ' is the linear functional on P(R) defined by '(p) = p(3).
Load more

Recommended publications
  • Bornologically Isomorphic Representations of Tensor Distributions
    Bornologically isomorphic representations of distributions on manifolds E. Nigsch Thursday 15th November, 2018 Abstract Distributional tensor fields can be regarded as multilinear mappings with distributional values or as (classical) tensor fields with distribu- tional coefficients. We show that the corresponding isomorphisms hold also in the bornological setting. 1 Introduction ′ ′ ′r s ′ Let D (M) := Γc(M, Vol(M)) and Ds (M) := Γc(M, Tr(M) ⊗ Vol(M)) be the strong duals of the space of compactly supported sections of the volume s bundle Vol(M) and of its tensor product with the tensor bundle Tr(M) over a manifold; these are the spaces of scalar and tensor distributions on M as defined in [?, ?]. A property of the space of tensor distributions which is fundamental in distributional geometry is given by the C∞(M)-module isomorphisms ′r ∼ s ′ ∼ r ′ Ds (M) = LC∞(M)(Tr (M), D (M)) = Ts (M) ⊗C∞(M) D (M) (1) (cf. [?, Theorem 3.1.12 and Corollary 3.1.15]) where C∞(M) is the space of smooth functions on M. In[?] a space of Colombeau-type nonlinear generalized tensor fields was constructed. This involved handling smooth functions (in the sense of convenient calculus as developed in [?]) in par- arXiv:1105.1642v1 [math.FA] 9 May 2011 ∞ r ′ ticular on the C (M)-module tensor products Ts (M) ⊗C∞(M) D (M) and Γ(E) ⊗C∞(M) Γ(F ), where Γ(E) denotes the space of smooth sections of a vector bundle E over M. In[?], however, only minor attention was paid to questions of topology on these tensor products.
    [Show full text]
  • Solution: the First Element of the Dual Basis Is the Linear Function Α
    Homework assignment 7 pp. 105 3 Exercise 2. Let B = f®1; ®2; ®3g be the basis for C defined by ®1 = (1; 0; ¡1) ®2 = (1; 1; 1) ®3 = (2; 2; 0): Find the dual basis of B. Solution: ¤ The first element of the dual basis is the linear function ®1 such ¤ ¤ ¤ that ®1(®1) = 1; ®1(®2) = 0 and ®1(®3) = 0. To describe such a function more explicitly we need to find its values on the standard basis vectors e1, e2 and e3. To do this express e1; e2; e3 through ®1; ®2; ®3 (refer to the solution of Exercise 1 pp. 54-55 from Homework 6). For each i = 1; 2; 3 you will find the numbers ai; bi; ci such that e1 = ai®1 + bi®2 + ci®3 (i.e. the coordinates of ei relative to the ¤ ¤ ¤ basis ®1; ®2; ®3). Then by linearity of ®1 we get that ®1(ei) = ai. Then ®2(ei) = bi, ¤ and ®3(ei) = ci. This is the answer. It can also be reformulated as follows. If P is the transition matrix from the standard basis e1; e2; e3 to ®1; ®2; ®3, i.e. ¡1 t (®1; ®2; ®3) = (e1; e2; e3)P , then (P ) is the transition matrix from the dual basis ¤ ¤ ¤ ¤ ¤ ¤ ¤ ¤ ¤ ¤ ¤ ¤ ¡1 t e1; e2; e3 to the dual basis ®1; ®2; a3, i.e. (®1; ®2; a3) = (e1; e2; e3)(P ) . Note that this problem is basically the change of coordinates problem: e.g. ¤ 3 the value of ®1 on the vector v 2 C is the first coordinate of v relative to the basis ®1; ®2; ®3.
    [Show full text]
  • Vectors and Dual Vectors
    Vectors By now you have a pretty good experience with \vectors". Usually, a vector is defined as a quantity that has a direction and a magnitude, such as a position vector, velocity vector, acceleration vector, etc. However, the notion of a vector has a considerably wider realm of applicability than these examples might suggest. The set of all real numbers forms a vector space, as does the set of all complex numbers. The set of functions on a set (e.g., functions of one variable, f(x)) form a vector space. Solutions of linear homogeneous equations form a vector space. We begin by giving the abstract rules for forming a space of vectors, also known as a vector space. A vector space V is a set equipped with an operation of \addition" and an additive identity. The elements of the set are called vectors, which we shall denote as ~u, ~v, ~w, etc. For now, you can think of them as position vectors in order to keep yourself sane. Addition, is an operation in which two vectors, say ~u and ~v, can be combined to make another vector, say, ~w. We denote this operation by the symbol \+": ~u + ~v = ~w: (1) Do not be fooled by this simple notation. The \addition" of vectors may be quite a different operation than ordinary arithmetic addition. For example, if we view position vectors in the x-y plane as \arrows" drawn from the origin, the addition of vectors is defined by the parallelogram rule. Clearly this rule is quite different than ordinary \addition".
    [Show full text]
  • A Some Basic Rules of Tensor Calculus
    A Some Basic Rules of Tensor Calculus The tensor calculus is a powerful tool for the description of the fundamentals in con- tinuum mechanics and the derivation of the governing equations for applied prob- lems. In general, there are two possibilities for the representation of the tensors and the tensorial equations: – the direct (symbolic) notation and – the index (component) notation The direct notation operates with scalars, vectors and tensors as physical objects defined in the three dimensional space. A vector (first rank tensor) a is considered as a directed line segment rather than a triple of numbers (coordinates). A second rank tensor A is any finite sum of ordered vector pairs A = a b + ... +c d. The scalars, vectors and tensors are handled as invariant (independent⊗ from the choice⊗ of the coordinate system) objects. This is the reason for the use of the direct notation in the modern literature of mechanics and rheology, e.g. [29, 32, 49, 123, 131, 199, 246, 313, 334] among others. The index notation deals with components or coordinates of vectors and tensors. For a selected basis, e.g. gi, i = 1, 2, 3 one can write a = aig , A = aibj + ... + cidj g g i i ⊗ j Here the Einstein’s summation convention is used: in one expression the twice re- peated indices are summed up from 1 to 3, e.g. 3 3 k k ik ik a gk ∑ a gk, A bk ∑ A bk ≡ k=1 ≡ k=1 In the above examples k is a so-called dummy index. Within the index notation the basic operations with tensors are defined with respect to their coordinates, e.
    [Show full text]
  • Multilinear Algebra
    Appendix A Multilinear Algebra This chapter presents concepts from multilinear algebra based on the basic properties of finite dimensional vector spaces and linear maps. The primary aim of the chapter is to give a concise introduction to alternating tensors which are necessary to define differential forms on manifolds. Many of the stated definitions and propositions can be found in Lee [1], Chaps. 11, 12 and 14. Some definitions and propositions are complemented by short and simple examples. First, in Sect. A.1 dual and bidual vector spaces are discussed. Subsequently, in Sects. A.2–A.4, tensors and alternating tensors together with operations such as the tensor and wedge product are introduced. Lastly, in Sect. A.5, the concepts which are necessary to introduce the wedge product are summarized in eight steps. A.1 The Dual Space Let V be a real vector space of finite dimension dim V = n.Let(e1,...,en) be a basis of V . Then every v ∈ V can be uniquely represented as a linear combination i v = v ei , (A.1) where summation convention over repeated indices is applied. The coefficients vi ∈ R arereferredtoascomponents of the vector v. Throughout the whole chapter, only finite dimensional real vector spaces, typically denoted by V , are treated. When not stated differently, summation convention is applied. Definition A.1 (Dual Space)Thedual space of V is the set of real-valued linear functionals ∗ V := {ω : V → R : ω linear} . (A.2) The elements of the dual space V ∗ are called linear forms on V . © Springer International Publishing Switzerland 2015 123 S.R.
    [Show full text]
  • Recent Developments in the Theory of Duality in Locally Convex Vector Spaces
    [ VOLUME 6 I ISSUE 2 I APRIL– JUNE 2019] E ISSN 2348 –1269, PRINT ISSN 2349-5138 RECENT DEVELOPMENTS IN THE THEORY OF DUALITY IN LOCALLY CONVEX VECTOR SPACES CHETNA KUMARI1 & RABISH KUMAR2* 1Research Scholar, University Department of Mathematics, B. R. A. Bihar University, Muzaffarpur 2*Research Scholar, University Department of Mathematics T. M. B. University, Bhagalpur Received: February 19, 2019 Accepted: April 01, 2019 ABSTRACT: : The present paper concerned with vector spaces over the real field: the passage to complex spaces offers no difficulty. We shall assume that the definition and properties of convex sets are known. A locally convex space is a topological vector space in which there is a fundamental system of neighborhoods of 0 which are convex; these neighborhoods can always be supposed to be symmetric and absorbing. Key Words: LOCALLY CONVEX SPACES We shall be exclusively concerned with vector spaces over the real field: the passage to complex spaces offers no difficulty. We shall assume that the definition and properties of convex sets are known. A convex set A in a vector space E is symmetric if —A=A; then 0ЄA if A is not empty. A convex set A is absorbing if for every X≠0 in E), there exists a number α≠0 such that λxЄA for |λ| ≤ α ; this implies that A generates E. A locally convex space is a topological vector space in which there is a fundamental system of neighborhoods of 0 which are convex; these neighborhoods can always be supposed to be symmetric and absorbing. Conversely, if any filter base is given on a vector space E, and consists of convex, symmetric, and absorbing sets, then it defines one and only one topology on E for which x+y and λx are continuous functions of both their arguments.
    [Show full text]
  • Residues, Duality, and the Fundamental Class of a Scheme-Map
    RESIDUES, DUALITY, AND THE FUNDAMENTAL CLASS OF A SCHEME-MAP JOSEPH LIPMAN Abstract. The duality theory of coherent sheaves on algebraic vari- eties goes back to Roch's half of the Riemann-Roch theorem for Riemann surfaces (1870s). In the 1950s, it grew into Serre duality on normal projective varieties; and shortly thereafter, into Grothendieck duality for arbitrary varieties and more generally, maps of noetherian schemes. This theory has found many applications in geometry and commutative algebra. We will sketch the theory in the reasonably accessible context of a variety V over a perfect field k, emphasizing the role of differential forms, as expressed locally via residues and globally via the fundamental class of V=k. (These notions will be explained.) As time permits, we will indicate some connections with Hochschild homology, and generalizations to maps of noetherian (formal) schemes. Even 50 years after the inception of Grothendieck's theory, some of these generalizations remain to be worked out. Contents Introduction1 1. Riemann-Roch and duality on curves2 2. Regular differentials on algebraic varieties4 3. Higher-dimensional residues6 4. Residues, integrals and duality7 5. Closing remarks8 References9 Introduction This talk, aimed at graduate students, will be about the duality theory of coherent sheaves on algebraic varieties, and a bit about its massive|and still continuing|development into Grothendieck duality theory, with a few indications of applications. The prerequisite is some understanding of what is meant by cohomology, both global and local, of such sheaves. Date: May 15, 2011. 2000 Mathematics Subject Classification. Primary 14F99. Key words and phrases. Hochschild homology, Grothendieck duality, fundamental class.
    [Show full text]
  • Basis (Topology), Basis (Vector Space), Ck, , Change of Variables (Integration), Closed, Closed Form, Closure, Coboundary, Cobou
    Index basis (topology), Ôç injective, ç basis (vector space), ç interior, Ôò isomorphic, ¥ ∞ C , Ôý, ÔÔ isomorphism, ç Ck, Ôý, ÔÔ change of variables (integration), ÔÔ kernel, ç closed, Ôò closed form, Þ linear, ç closure, Ôò linear combination, ç coboundary, Þ linearly independent, ç coboundary map, Þ nullspace, ç cochain complex, Þ cochain homotopic, open cover, Ôò cochain homotopy, open set, Ôý cochain map, Þ cocycle, Þ partial derivative, Ôý cohomology, Þ compact, Ôò quotient space, â component function, À range, ç continuous, Ôç rank-nullity theorem, coordinate function, Ôý relative topology, Ôò dierential, Þ resolution, ¥ dimension, ç resolution of the identity, À direct sum, second-countable, Ôç directional derivative, Ôý short exact sequence, discrete topology, Ôò smooth homotopy, dual basis, À span, ç dual space, À standard topology on Rn, Ôò exact form, Þ subspace, ç exact sequence, ¥ subspace topology, Ôò support, Ô¥ Hausdor, Ôç surjective, ç homotopic, symmetry of partial derivatives, ÔÔ homotopy, topological space, Ôò image, ç topology, Ôò induced topology, Ôò total derivative, ÔÔ Ô trivial topology, Ôò vector space, ç well-dened, â ò Glossary Linear Algebra Denition. A real vector space V is a set equipped with an addition operation , and a scalar (R) multiplication operation satisfying the usual identities. + Denition. A subset W ⋅ V is a subspace if W is a vector space under the same operations as V. Denition. A linear combination⊂ of a subset S V is a sum of the form a v, where each a , and only nitely many of the a are nonzero. v v R ⊂ v v∈S v∈S ⋅ ∈ { } Denition.Qe span of a subset S V is the subspace span S of linear combinations of S.
    [Show full text]
  • SERRE DUALITY and APPLICATIONS Contents 1
    SERRE DUALITY AND APPLICATIONS JUN HOU FUNG Abstract. We carefully develop the theory of Serre duality and dualizing sheaves. We differ from the approach in [12] in that the use of spectral se- quences and the Yoneda pairing are emphasized to put the proofs in a more systematic framework. As applications of the theory, we discuss the Riemann- Roch theorem for curves and Bott's theorem in representation theory (follow- ing [8]) using the algebraic-geometric machinery presented. Contents 1. Prerequisites 1 1.1. A crash course in sheaves and schemes 2 2. Serre duality theory 5 2.1. The cohomology of projective space 6 2.2. Twisted sheaves 9 2.3. The Yoneda pairing 10 2.4. Proof of theorem 2.1 12 2.5. The Grothendieck spectral sequence 13 2.6. Towards Grothendieck duality: dualizing sheaves 16 3. The Riemann-Roch theorem for curves 22 4. Bott's theorem 24 4.1. Statement and proof 24 4.2. Some facts from algebraic geometry 29 4.3. Proof of theorem 4.5 33 Acknowledgments 34 References 35 1. Prerequisites Studying algebraic geometry from the modern perspective now requires a some- what substantial background in commutative and homological algebra, and it would be impractical to go through the many definitions and theorems in a short paper. In any case, I can do no better than the usual treatments of the various subjects, for which references are provided in the bibliography. To a first approximation, this paper can be read in conjunction with chapter III of Hartshorne [12]. Also in the bibliography are a couple of texts on algebraic groups and Lie theory which may be beneficial to understanding the latter parts of this paper.
    [Show full text]
  • Homotopy Theory and Generalized Duality for Spectral Sheaves
    IMRN International Mathematics Research Notices 1996, No. 20 Homotopy Theory and Generalized Duality for Spectral Sheaves Jonathan Block and Andrey Lazarev We would like to express our sadness at the passing of Robert Thomason whose ideas have had a lasting impact on our work. The mathematical community is sorely impoverished by the loss. We dedicate this paper to his memory. 1 Introduction In this paper we announce a Verdier-type duality theorem for sheaves of spectra on a topological space X. Along the way we are led to develop the homotopy theory and stable homotopy theory of spectral sheaves. Such theories have been worked out in the past, most notably by [Br], [BG], [T], and [J]. But for our purposes these theories are inappropri- ate. They also have not been developed as fully as they are here. The previous authors did not have at their disposal the especially good categories of spectra constructed in [EKM], which allow one to do all of homological algebra in the category of spectra. Because we want to work in one of these categories of spectra, we are led to consider sheaves of spaces (as opposed to simplicial sets), and this gives rise to some additional technical difficulties. As an application we compute an example from geometric topology of stratified spaces. In future work we intend to apply our theory to equivariant K-theory. 2 Spectral sheaves There are numerous categories of spectra in existence and our theory can be developed in a number of them. Received 4 June 1996. Revision received 9 August 1996.
    [Show full text]
  • Duality for Mixed-Integer Linear Programs
    Duality for Mixed-Integer Linear Programs M. Guzelsoy∗ T. K. Ralphs† Original May, 2006 Revised April, 2007 Abstract The theory of duality for linear programs is well-developed and has been successful in advancing both the theory and practice of linear programming. In principle, much of this broad framework can be ex- tended to mixed-integer linear programs, but this has proven difficult, in part because duality theory does not integrate well with current computational practice. This paper surveys what is known about duality for integer programs and offers some minor extensions, with an eye towards developing a more practical framework. Keywords: Duality, Mixed-Integer Linear Programming, Value Function, Branch and Cut. 1 Introduction Duality has long been a central tool in the development of both optimization theory and methodology. The study of duality has led to efficient procedures for computing bounds, is central to our ability to perform post facto solution analysis, is the basis for procedures such as column generation and reduced cost fixing, and has yielded optimality conditions that can be used as the basis for “warm starting” techniques. Such procedures are useful both in cases where the input data are subject to fluctuation after the solution procedure has been initiated and in applications for which the solution of a series of closely-related instances is required. This is the case for a variety of integer optimization algorithms, including decomposition algorithms, parametric and stochastic programming algorithms, multi-objective optimization algorithms, and algorithms for analyzing infeasible mathematical models. The theory of duality for linear programs (LPs) is well-developed and has been extremely successful in contributing to both theory and practice.
    [Show full text]
  • Duality in MIP
    Duality in MIP Generating Dual Price Functions Using Branch-and-Cut Elena V. Pachkova Abstract This presentation treats duality in Mixed Integer Programming (MIP in short). A dual of a MIP problem includes a dual price function F , that plays the same role as the dual variables in Linear Programming (LP in the following). The price function is generated while solving the primal problem. However, di®erent to the LP dual variables, the characteristics of the dual price function depend on the algorithmic approach used to solve the MIP problem. Thus, the cutting plane approach provides non- decreasing and superadditive price functions while branch-and-bound algorithm generates piecewise linear, nondecreasing and convex price functions. Here a hybrid algorithm based on branch-and-cut is investigated, and a price function for that algorithm is established. This price function presents a generalization of the dual price functions obtained by either the cutting plane or the branch-and-bound method. 1 Introduction Duality in mathematical programming is used in a variety of applications. Apart from con- ceptual interest it provides interesting economic interpretations of the problem. Moreover, using dual information usually improves the performance of an algorithm. Thus, there exist many results on duality in linear programming (LP) (e.g. see Gass (1985)). Results on duality in integer programming (IP) also exist (Wolsey (1981)). While algorithms for LP produce unique dual programs (apart from degenerating programs), that are relatively easy to obtain, IP algorithms generate a dual function whose characteristics depend on the method used to solve the primal IP problem. Wolsey (1981) characterized this function for 73 the branch-and-bound and the cutting plane methods.
    [Show full text]