[Bi Brict Tabie of Contents B Tobie of Contents. 14+ The Matrix Equation Ax: Answers to Checkpoints: 19 1, False. See the Warning after Theorem 4. If you missed this, you are not studying the text properly. You should read the text thoroughly before you look at the Study Guide and before ‘you work on the exercises. 2QLetAsly v, w= ‘ f, | Row reduce the augmented matrix for by Ax=b to determine values of b;,. . . , bs that make the equation inconsistent. 101 4) 71 0 1 6 1oO 1 8 o 10 4/|0 10 & o10 6 -1 0 0 &| |o 1 +8) |o 1 +b, 0 -1 ba -1 b& 0 0 -1 b+, 4 b +h 00 0 by +b, +h +b, 0 0 0 0 1 0 Take b = (1, 1, 0, 0), for example, or any other choice of by, . .. , bs whose sum is not zero. Mastering Linear Algebra Concepts: Span Please begin by reviewing “How to Study Linear Algebra,” atthe beginning of this Study Guide. To really understand a key concept, you need to form an image in your mind that consists of the basic definition(s) together with many related ideas. Your goal at this point is to collect various ideas associated with the set Span{v, .. . vp} and the concept of a set that “spans” R". Here are specific things to do now as you prepare a sheet (or sheets) for review and reference. + Write the definition of Span, ..., vp}. (Leam it word for word.) + Write the definition of the phrase: (vi. .... vp} spans R". (See page 43.) Here span is a verb rather than a noun as in Span{vj, .. ¥p}. ‘Add the equivalent description (not definition) of what is meant for a vector b to be in Span{v;, ....¥p}- (See page 35.) Copy Theorem 4 word for word. (If you try to rephrase or summarize it in your own words, you are likely to change the meaning.) Copyright © 2006 Pearson Aakdson-Wesley, Allright reserved.EB Brief Tabio of Contents|] {B Table of Contents CHAPTER 1 + Linear Equations in Linear Algebra ‘Sketch some geometric interpretations of Span( vi, ..., vp]. (Select some of Figs. 8, 10, 11, and Exercises 19, 20 in Section 1.3.) Identify special cases. (Describe Span(u} and Span(u, ¥).) ‘Summarize algorithms or typical computations (such as Example 6 and Exercises 11-14, 17, 18, 25, and 26 in Section 1.3, or Example 3 and Exercises 13-22 in Section 1.4. Describe connections with other concepts. (See pages 42-43.) Whenever you encounter new examples or situations that help you understand the concept of a spanning set, add them to this review sheet. MATLAB gauss and bgauss To solve Ax = b, row reduce the matrix M = [A b]. The command x = [5;3; - 7) creates a column vector x with entries 5, 3,7. Matrix vector multiplication is A*x. To speed up row reduction of M = [A bl, the command gauss(M,r) will use the leading entry in row r of M as a pivot, and use row replacements to create zeros in the pivot column below this pivot entry. ‘The result is stored in the default matrix “ans”, unless ‘you assign the result to some other variable, such as M itself. For the backward phase of row reduction, use bgauss (M,r), which selects the leading entry in row r of M as the pivot, and creates zeros in the column above the pivot. Use scale to create leading I's in the pivot positions. The commands gauss, gauss, and scale are in the Laydata Toolbox, which you can download from the web. 1.5 | SOLUTION SETS OF LINEAR SYSTEMS Many of the concepts and computations in linear algebra involve sets of vectors which are visualized geometrically as lines and planes. ‘The most important examples of such sets are the solution sets of linear systems, KEY IDEAS ‘Visualize the solution set of a homogeneous equation Ax = 0 as: + the single point 0, when Ax = 0 has only the trivial solution, + line through 0, when Ax =0 has one free variable, + aplane through 0, when Ax = 0 has two free variables. (For more than two free variables, also use a plane through 0.) Copyright © 2006 Pearson Aakdson-Wesley, Allright reserved.EB Brief Tabio of Contents|] {B Table of Contents + Solution Sets of Linear Equations 1-21 For b # 0, visualize the solution set of Ax = b as: + empty, if b is not a linear combination of the columns of A, ‘+ one nonzero point (vector), when Ax = b has a unique solution, + alline not through 0, when Ax = b is consistent and has one free variable, + aplane not through 0, when Ax = b is consistent and has two or more free variables. ‘The solution set of Ax = b is said to be described implicitly, because the equation is a condition an x must satisfy in order to be in the set, yet the equation does not show how to find such an x. When the solution set of Ax = 0 is written as Span{vi, ... , Vp}, the set is said to be described explicitly, each element in the set is produced by forming a linear combination of v1, ..., Vp ‘A common explicit description of a set is an equation in parametric vector form. Examples are: x= tv, ane through 0 in the direction of v, x=pttv, Aline trough p in the direction of v, x = xu + 2Gv, aplane through 0, u, and v, X =P +X: + 25V, a plane through p parallel othe plane whose equation is x = +.5v. An equation in parametric vector form describes a set explicitly because the equation shows how to produce each x in the set. To solve an equation Ax = b means to find an explicit description of the solution set. If the system is inconsistent, the solution set is empty. Otherwise, the description of all solutions can be written in parametric vector form, in which the parameters are the free variables from the system. Important: The number of free variables in Ax = b depends only on A, not on b. Theorem 6 and the paragraph following it are important. They describe how the solutions of are related when the solution set of Ax = b is nonempty. See Figs. 5 and 6. 5-16, 29-32, 37. ‘SOLUTIONS TO EXERCISES 1, Reduce the augmented matrix to echelon form and circle the pivot positions. If a column of the coefficient matrix is not a pivot column, the corresponding variable is free and the system of equations has a nontrivial solution. Otherwise, the system has only the trivial solution. 2-5 8 Oo) [2 -5 8 0 58 0 2 7 1 of-lo 12 9 o|-l0 G3 9 0 4 27 0} {0 2-9 of [0 00 0 ‘The variable x is free, so the system has a nontrivial solution, Copyright © 2006 Pearson Aakdson-Wesley, Allright reserved.Briof Table of Contents|] 1 Table of Contents 1-22 CHAPTER + Linear Equations in Linear Algebra 7. Always use the reduced echelon form of an augmented matrix to find the solutions of a system. See the text's discussion of back = ‘on pages 22-23. if 3 i: 9 +} + 9x, = -8 on x0) 4 Os 4x = If you wrote at like the system. Ie then you made a common mistake. The matrix in the text problem is a coefficient matrix, not an augmented matrix. You should row reduce [A 0}. The correct system of equations is, @_ +95 - 8% =0 @- 4% + 5x, = 0 ‘The basic variables are x; and xs, with x3 and x, free. Next, x; =—91s + 8x, and y= 4 — Sty ‘The general solution is ~94,+8x,] [-9,] [ 8x, aJa| 49-58 |] 4u |.) Se | xy x % 0 * my 0 % ‘The solution set is the same as Span(u, v}, where u = (-9, 4, 1, 0) and v = (8, -5,0, 1). Originally, the solution set was described implicitly, by a set of equations. Now the solution set is described explicitly, in parametric vector form. 3-5 0] fi 4 ~ 0 -1 0) Jo 1-4 0} Jo 0 0 0} lo + 5x, =0 - 420 = 4x, = 0 o=0 4 0 0 0 Copyright © 2006 Pearson Aakdson-Wesley, Allright reserved.EB Brief Tabio of Contents|] {B Table of Contents + Solution Sets of Linear Equations 1-23, ‘Some students are not sure what to do with x,. Some ignore it; others set it equal to zero. In fact, x4 is free; there is no constraint on x, at all. The basic variables are x), 15, ands. The remaining variables are free. So, x; = 4x3 — Sx, x3 = 4, and x5 = 4xg, with xp, x4, and x. free. In parametric vector form, 4x, -5% Sx, 0 0 0 % % The solution set is the same as Span{u, v, w). Study Tip: When solving a system, identify (and perhaps circle) the basic variables. All other variables are free. 13. To write the general solution in parametric vector form, pull out the constant terms that do not involve the free variable: x] [5+4x 5) [ 4u] [5 4 x=] x, |=|-2-7x, |=|-2]+]-7x, |=|-2]+2,] -7 4, % 0 4) | 0 1 t ® ape P q 5 4 Geometrically, the solution set is the line through | ~2 parallel to | ~7 |. 0 1 Checkpoint: Let A be a 2 2 matrix. Answer True or False: If the solution set of Ax = 0 is a line through the origin in RR? and if b + 0, then the solution set of Ax = b is a line not through the origin. 19. The line through a parallel to b can be written as x = a + tb, where r represents a parameter: 4) [2], [3 2-51 x=|* |=] lee) >] or a} Lol” 3} ly =3 23, a. See the first paragraph of the subsection titled “Homogeneous Linear Systems.” . See the first two sentences of the subsection titled “Parametric Vector Form.” Copyright © 2006 Pearson Aakdson-Wesley, Allright reserved.EB Brief Tabio of Contents|] {B Table of Contents 1.7 + Linear Independence 1-31 7. Study the equation Ax = 0. Some people may start with the method of Example 2: 1 4-3 oo0 14300 43 00 2-7 5 1 Oj-|o 1-1 1 O|-]o@-1 1 0 ri SF 7 -# F ol -5 5 0 oo ©-~ 0. But this is a waste of time. There are only 3 rows, so there are at most three pivot positions. Hence, at least one of the four variables must be free. So the equation Ax = 0 has a nontrivial solution and the columns of A are linearly dependent. Warning: Exercise 9 and Practice Problem 3 emphasize that to check whether a set such as {¥1, v2, V5) is linearly dependent, it is not wise to check instead whether v; is a linear combination of vi and vs. 13. To study the linear esate of three vectors, say vi, ¥2. V3, fow reduce the augmented matrix tn ‘ * :3 oF fe baci 3 ‘The equation xi¥; +.2¥2 + Xs¥s = 0 has a free variable and hence a nontrivial solution no matter what the value of h. So the vectors are linearly dependent for all values of h. Checkpoint: What is wrong with the following statement? ‘The vectors (MESH. are linearly dependent “because there is a free variable,” or “because there are more variables than equations.” 18s 13)", 7. are more vectors (4) than entries in the vectors. On atest, you probably will not have to know the theorem number. Check with your instructor. 19. The set is linearly independent because neither vector is a multiple of the other vector. [Two of the entries in the first vector are —4 times the corresponding entry in the second vector. But this multiple does not work for the third entries.] 15, The set {GIG}: obviously linearly dependent, by Theorem 8, because thee Copyright © 2006 Pearson Aakdson-Wesley, Allright reserved.EB Brief Tabio of Contents|] {B Table of Contents ‘Chapter 1 Supplementary Exercises / Glossary Checklist 1-47 o* 13, The reduced echelon form of A looks like E=|0 1 *|. Since Eis row equivalent to A, 00 3] fo -2|=]0]. By 1 0 10 the equation Ex = 0 has the same solutions as Ax=0. Thus |Q 1 00 0 10-3 inspection, E=|0 1 2 00 0 17. Here are two arguments. ‘The first is a “direct” proof. The second is called a “proof by contradiction.” i. Since {v4, vz, ¥s) isa linearly independent set, v, #0. Also, Theorem 7 shows that ¥2 cannot be a multiple of vj, and ¥; cannot be a linear combination of ¥, and ¥:. By hypothesis, vs is not a linear combination of ¥;, v2, and vs. Thus, by Theorem 7, vj, ¥2, ¥s, Ve} cannot be a linearly dependent set and so must be linearly independent. . Since (Vy, V2, Ys) is a linearly independent set, v; #0. Suppose that (¥), V2, Vi Va) iS linearly dependent. Then, by Theorem 7, one of the vectors in the set is a linear ‘combination of the preceding vectors. This vector cannot be vs because vg is not in Span{vi, ¥2, V3}. Also, none of the vectors in {¥;, ¥a, ¥s} is a linear combination of the preceding vectors, by Theorem 7. So the linear dependence of (vi, V2, ¥5, Vs) is impossible, and {V), Vzy V3. Vs} is linearly independent. Chapter 1 GLOSSARY CHECKLIST Check your knowledge by attempting to write definitions of the terms below. Then compare your ‘work with the definitions given in the text’s Glossary. Ask your instructor which definitions, if any, might appear on a test. affine transformation: A mapping T: R’ +R” of the form T(x) augmented matrix: A matrix made up of a. back-substitution (with matrix notation): The . . . phase of row reduction of an . basic variable: A variable in a linear system that... codomain (of T: R"R"): The set . . . that contains . .. . coefficient matrix: A matrix whose entries are... . Copyright © 2006 Pearson Aakdson-Wesley, Allright reserved.EB Brief Tabio of Contents|] {B Table of Contents 23. + Characterization of Inverible Matrices. 2-9 ‘To show that AB = I, it suffices to show that Ab, = ¢; for each j. For j 2qhindy ACG; — @jo1) = AG) — A@ pes = 8) — Aye, = and Ab, = Ae, =a, =e). Next, observe that a,=e) +++ +e, for each j. Thus, Ba, = Ble; +--+ e,) =bj+-+- +b, = (6) — ee) + (Gri — Opa) #2 + Ger On) HOD =E) This proves that BA = /. Combined with the first part, this proves that Li : a is the only one whose entries are 1,-I, and 0. With only three possibilities for each entry, the construction of C can be done by trial and error. This is probably faster than setting up a system of 4 equations in 6 unknowns. The fact that A cannot be invertible follows from Exercise 25 in Section 2.1, because A is not square. 37, There are many possibilities for C, but C = [ MATLAB Constructing A“ IfA is a 5x5 matrix, then the command M = [A eye(5)] creates the augmented matrix [A 1). Use ga swap, bgauss, and scale toreduce [A 1]. See page 1-17. MATLAB has other commands that row reduce matrices, invert matrices, and solve equations Ax = b. They will be introduced later, after you have studied the concepts and algorithms in this section. 2.3 CHARACTERIZATION OF INVERTIBLE MATRICES In many linear algebra texts, the equivalent of Chapter 4 is extremely difficult for students. But ‘you won't have problems if you prepare well now, because you are already leaming basic ideas that will be presented again in Chapter 4. Review the major concepts from the previous sections, and plan for more study time here than you ordinarily spend on one section. KEY IDEAS. ‘The Invertible Matrix Theorem (IMT) only applies to square matrices. However, some groups of these statements in the IMT are also equivalent for rectangular matrices. The following table will help you remember other important theorems as well as the IMT. (See Theorem 4 in Section 1.4, Theorems 11 and 12 in Section 1.9, Theorem 5 in Section 2.2, and Theorem 9 in Section 2.3.) All of the statements in the table are equivalent when A is square (m =n = p). Copyright © 2006 Pearson Aakdson-Wesley, Allright reserved.[Bi Brit Tabie of Contents B Tobie of Contents. 224 CHAPTER? + Matrix Algebra Appendix: Permuted LU Factorizations Any mXn matrix A admits a factorization A = LU, with U in echelon form and L a permuted unit lower triangular matrix. That is, L is a matrix such that a permutation (rearrangement) of its rows (using row interchanges) will produce a lower triangular matrix with I's on the diagonal. ‘The construction of L and U, illustrated below, depends on first using row replacements to reduce A to a permuted echelon form V and then using row interchanges to reduce V to an echelon form U. By watching the reduction of 4 to V, we can easily construct a permuted unit lower triangular matrix L with the property that the sequence of operations changing A into U also changes L into /. This property will guarantee that A= LU. (See the paragraph before Example 2 in the text.) The following algorithm reduces any matrix to a permuted echelon form. In the algorithm when a row is covered, we ignore it in later calculations 1. Begin with the leftmost nonzero column. Choose any nonzero entry as the pivot. Designate the corresponding row as a pivot row. Use row replacements to create zeros above and below the pivot (in all uncovered rows), Then cover that pivot row. Repeat steps 1 and 2 on the uncovered submatrix, if any, until all nonzero entries are covered. This algorithm forces each pivot to be to the right of the preceding pivots; when the rows are rearranged with the pivots in stair-step fashion, all entries below each pivot will be zero. ‘Thus, the algorithm produces a permuted echelon matrix. Whenever a pivot is selected, the column containing the pivot will be used to construct a column of L, as we shall see. ‘As an example, choose any entry in the first column of the following matrix as the first pivot, and use the pivot to create zeros in the rest of column 1. We choose the (3, 1)-entry. CT call his column & 36 -8 -5 93 OER] <— tst pivot row “5 7. Copyright © 2006 Pearson Aakdson-Wesley, Allright reserved.