Skip to main content

2012 | Buch

The Linear Algebra a Beginning Graduate Student Ought to Know

insite
SUCHEN

Über dieses Buch

Linear algebra is a living, active branch of mathematics which is central to almost all other areas of mathematics, both pure and applied, as well as to computer science, to the physical, biological, and social sciences, and to engineering. It encompasses an extensive corpus of theoretical results as well as a large and rapidly-growing body of computational techniques. Unfortunately, in the past decade, the content of linear algebra courses required to complete an undergraduate degree in mathematics has been depleted to the extent that they fail to provide a sufficient theoretical or computational background. Students are not only less able to formulate or even follow mathematical proofs, they are also less able to understand the mathematics of the numerical algorithms they need for applications. Certainly, the material presented in the average undergraduate course is insufficient for graduate study. This book is intended to fill the gap which has developed by providing enough theoretical and computational material to allow the advanced undergraduate or beginning graduate student to overcome this deficiency and be able to work independently or in advanced courses. The book is intended to be used either as a self-study guide, a textbook for a course in advanced linear algebra, or as a reference book. It is also designed to prepare a student for the linear algebra portion of prelim exams or PhD qualifying exams. The volume is self-contained to the extent that it does not assume any previous formal knowledge of linear algebra, though the reader is assumed to have been exposed, at least informally, to some of the basic ideas and techniques, such as manipulation of small matrices and the solution of small systems of linear equations over the real numbers. More importantly, it assumes a seriousness of purpose, considerable motivation, and a modicum of mathematical sophistication on the part of the reader. In the latest edition, new major theorems have been added, as well as many new examples. There are over 130 additional exercises and many of the previous exercises have been revised or rewritten. In addition, a large number of additional biographical notes and thumbnail portraits of mathematicians have been included.

Inhaltsverzeichnis

Frontmatter
1. Notation and Terminology
Abstract
The set-theoretic and functional notation and terminology which will be used throughout the book is summarized in this short chapter.
Jonathan S. Golan
2. Fields
Abstract
Abstract fields, of zero and nonzero characteristic, are defined and their elementary properties are proven. Several important examples are discussed.
Jonathan S. Golan
3. Vector Spaces Over a Field
Abstract
In this chapter, abstract vector spaces and their subspaces are introduced and their elementary properties are established. Elementary properties of the lattice of subspaces of a vector space, such as the Modular Law, are proven. Direct products and coproducts are used to construct new spaces from given ones. Generating sets for vector spaces are defined and studied, and the notion of a finitely-generated space is introduced.
Jonathan S. Golan
4. Algebras Over a Field
Abstract
Algebras over a field, both associative and nonassociative, are introduced and many examples are given, among them Lie algebras and Jordan algebras. Polynomial algebras are studied in detail and results such as the division algorithm for polynomials are established. Karatsuba’s algorithm is introduced as an example of faster polynomial multiplication. Reducibility and factorization of polynomials are considered, as are the Fundamental Theorem of Algebra and its implications.
Jonathan S. Golan
5. Linear Independence and Dimension
Abstract
The notions of linear independence and bases are defined and studied for arbitrary vector spaces. The notion of dimension is defined. To study bases for arbitrary vector spaces, the Hausdorff Maximum Principle is introduced and used. The properties of finite-dimensional vector spaces are considered. Finally, independence and complements in the lattice of subspaces of a vector space are studied. Among the examples given are the quaternion algebras, Hamel bases, and the complexification of real vector spaces.
Jonathan S. Golan
6. Linear Transformations
Abstract
The notions of linear transformations, monomorphisms, epimorphisms, and isomorphisms between vector spaces and algebras over a field are introduced and studied, as are the notions of the kernel and image of a linear transformation. The rank and nullity of a linear transformation are defined and their properties are proven. The importance of the fact that the action of a linear transformation is completely determined by its values on any basis is emphasized through several examples.
Jonathan S. Golan
7. The Endomorphism Algebra of a Vector Space
Abstract
Endomorphisms and automorphisms of vector spaces and algebras over a field are introduced and the notion of the endomorphism algebra of a vector space is explored. The importance of idempotent elements of this algebra (namely, projections) is emphasized. The group of automorphisms is also considered. The notion of invariance of a subspace under an endomorphism is introduced.
Jonathan S. Golan
8. Representation of Linear Transformations by Matrices
Abstract
The representation of linear transformations between finitely-generated vector spaces by matrices is studied. As a consequence of this, the product of matrices of appropriate sizes is defined and its properties are studied.
Jonathan S. Golan
9. The Algebra of Square Matrices
Abstract
The algebra of square matrices with entries in a fixed associate algebra over a field are considered. Special classes of square matrices (diagonal, tridiagonal, upper triangular, symmetric, Vandermonde, etc.) are defined and their properties are identified. Nonsingular matrices are studied in detail and methods for identifying them and computing matrix inverses are presented. Computational methods for matrix manipulation, including the Strassen–Winograd algorithm and LU-decompositions are discussed.
Jonathan S. Golan
10. Systems of Linear Equations
Abstract
Direct and iterative methods for solution of systems of linear equations are studied, including the Gaussian elimination, Jacobi iteration, and Gauss–Seidel iteration. Overrelaxation methods are described. Numerical considerations are introduced, including condition numbers.
Jonathan S. Golan
11. Determinants
Abstract
Determinants of square matrices are defined and their properties are studied. Techniques for calculating determinants over various fields are introduced. Adjoints of square matrices are introduced and their use in computing the inverse of a nonsingular matrix are proven. Cramer’s Theorem is established. Padé approximants are defined using determinants.
Jonathan S. Golan
12. Eigenvalues and Eigenvectors
Abstract
Eigenvalues and the associated eigenvectors of an endomorphism of a vector space are defined and studied, as is the spectrum of an endomorphism. The characteristic polynomial of a matrix is considered and used to define the characteristic polynomial of the endomorphism of a finitely-generated vector space. This leads to the notions of geometric and algebraic multiplicities of eigenvalues. The minimal polynomial is then defined and studied. Among the theorems proven are the Cayley–Hamilton theorem and Burnside’s theorem. Von Mises’ algorithm for the computation of the dominant eigenvalue is introduced as an example of a iterative algorithm for eigenvalue computation.
Jonathan S. Golan
13. Krylov Subspaces
Abstract
Krylov subspaces are studied theoretically and as the foundation of Krylov iterative algorithms for approximating the solutions to systems of linear equations. The Rational Decomposition Theorem for nilpotent endomorphisms is proven and used to define the Jordan canonical form. Every square matrix over an algebraically-closed field is shown to be a product of two symmetric matrices and to be similar to its transpose.
Jonathan S. Golan
14. The Dual Space
Abstract
Linear functionals and the dual space of a vector space are defined and characterized. Every vector space is shown to be canonically embeddable in its second dual. Maximal subspaces are characterized as kernels of nontrivial linear functionals. The trace of a square matrix is studied in detail. Over a field of characteristic 0, a square matrix is shown to have trace 0 if and only if it is the Lie product of two matrices. Taber’s theorem is established.
Jonathan S. Golan
15. Inner Product Spaces
Abstract
Real and complex inner product spaces are defined and several examples are studied. Elementary properties of inner products, such as the Cauchy–Schwarz–Bunyakovsky Theorem and Minkowski’s inequality are proven. The Lagrange identity relating inner and cross products in three-dimensional real vector spaces is proven. Normed spaces are defined and various examples of norms are considered, including spectral norms and various matrix norms. The Hahn–Banach Theorem, Gershgorin’s Theorem, and the Diagonal Dominance Theorem are proven. Matrix exponentials are studied.
Jonathan S. Golan
16. Orthogonality
Abstract
The notion of orthogonality in an inner product space is introduced and several examples are presented, including Legendre and Chebyshev polynomials. The Gram–Schmidt Theorem and Hadamard’s inequality are proven. Orthogonal complements of subspaces are introduced, as are orthogonal projections. Every finitely-generated inner product space is shown to have an orthonormal basis and the properties of orthonormal bases are studied. QR-decompositions are introduced. The Riesz Representation Theorem is proven for finitely-generated inner product spaces. The notion of the adjoint of a linear transformation between inner product spaces is introduced and studied.
Jonathan S. Golan
17. Selfadjoint Endomorphisms
Abstract
Selfadjoint endomorphisms of inner product spaces are defined and studied. Any selfadjoint endomorphism of a finitely-generated inner product space is shown to have a nonempty spectrum. The notion of orthogonal decomposition of an endomorphism is introduced. Selfadjoint endomorphisms of finitely-generated inner product spaces are shown to be orthogonally diagonalizable, with the converse true for spaces over the real numbers. Positive-definite endomorphisms are introduced and characterized. Application is made to Cholesky decompositions. Isometries of finitely-generated inner product spaces are studied and characterized.
Jonathan S. Golan
18. Unitary and Normal Endomorphisms
Abstract
Unitary endomorphisms of an inner product space are introduced and studied. In particular, unitary and orthogonal matrices are considered. Many examples are considered, among them Householder matrices and special orthogonal matrices. Unitarily-similar matrices are defined and Schur’s Theorem is proven. Normal endomorphisms of an inner product space are also considered and their properties proven, leading up to a proof of the Spectral Decomposition Theorem and the Singular Value Decomposition Theorem.
Jonathan S. Golan
19. Moore–Penrose Pseudoinverses
Abstract
Moore–Penrose pseudoinverses of linear transformations between inner product spaces are defined and their existence and uniqueness are established. In particular, the application of pseudoinverses of matrices to finding best approximations of solutions of systems of linear equations is considered and related to the least-squares method.
Jonathan S. Golan
20. Bilinear Transformations and Forms
Abstract
Bilinear transformations and bilinear forms are introduced and studied. Their matrix representation, and especially the representation of symmetric bilinear forms, is presented. Orthogonality relative to a bilinear form is considered. These notions are then used to define the tensor product of vector spaces, and the properties of the tensor product are considered in detail.
Jonathan S. Golan
Backmatter
Metadaten
Titel
The Linear Algebra a Beginning Graduate Student Ought to Know
verfasst von
Jonathan S. Golan
Copyright-Jahr
2012
Verlag
Springer Netherlands
Electronic ISBN
978-94-007-2636-9
Print ISBN
978-94-007-2635-2
DOI
https://doi.org/10.1007/978-94-007-2636-9

Premium Partner