Chapter 7: Eigenvectors and lines 7.1 Invariant points 2 Eigenvectors and eigenvalues The characteristic equation 7.4 Diagonalisation This chapter introduces the idea of studying points and lines which are fixed by matrix transformations. When you have completed it, you will be able to find the invariant points and lines of transformations: snvectors and eigenvalues of a matrix transformation; know what is meant by the understand why the ci nivalues are the roots of the characteristic equation be able to diagonalise square matrices which have a full set of eigenvectors 967.1 Invariant points and lines ant point of a transformation is a point which is unchanged by the transformation For example, a reflection in the line y= x leaves every point on the line y= unchanged variant points. left The invariant points of a matriy transformation can be found by solving a matrix equation The invariant points of the transformation with matrix M can be found by solving Mx = Example 7.1.1 Find the invariant points of the transformation with mattis. where £20. Solution SN the invariant points are al the points on the line An invariant line of a transformation is a line which is left ‘fixed’ by the transformation, In some cases, this line will consist of invariant points. In others, it will consist of points which are moved around on the line. For example, a reflection in the line y= x has infinitely many invariant lines. y invariant fink ~ a line of invariant pointsThe invariant lines of a matrix transformation can also be found by solving equations The invariant lines of the transformation with matrix Mcan be found by substituting x and Mx in the same equation of a line Example 7.1.2 Find, in the form y = mx ~c, the invariant lines of the transformation with matrix: Solution )_lmr+e) meso) | ox | > ven ie)-e (len x= (m= Tes forall x Put ¥=0, then (mi 1jo=0 m= Lore=0 Then (I-m )y=0. for ally => m= ~1 So, or m=le=0 Sys or yor The invariant points and lines of a transformation are closely connected with its geometrical significance. Thus a rotation in 3-dimensional space alway’ has a line of invariant points the axis of the rotation — but only has other invariant lines if the rotation is 180° For complicated transformations, it can be useful to determine any invariant points and lines before attempting to decide on the nature of the transformation. Section 7.2 develops these ideas furtherExercise 7A. |. Find all invariant lines, of the form y= mx, for the matrix transformations -5 2. The plane transformation TT is defined by (a) Show that Phas a line of invariant points and find its equation (b) Show that there is an infinite number of invariant lines and find the general equation of such lines. 1404) 3. (a) Given the matrix M evaluate MP and the determinant of M. ] [: ] fu (b) Fina vestr| “such that Mi | [ | and alsoa veetor |" such that y vy Ly uw _[-w w [ | (c) Describe, in geometrical terms, the transformation represented by the matrix M. (JMB, 1977] yoandM= 3 6 4 z | 5 (a) Prove that T vansforms the plane x ine is independent of the value of vinto a line, and show that the direction of this (b) Given that this line passes through the point (5, -S, 15), determine the value of UMB. 1979] 4 12 5. Find a line of invariant points for the transformation with matrix M=| 26 6 8 -6 -3) 997.2 Eigenvectors and eigenvalues Ifa square matrix Mand vector y satisfy the equation Mv-av. v0. then vis called an eigenvector of the transformation and Ais called an eigenvalue, Itvis an eigenvector, then the line any multiple of vi vis an invariant line through the origin, In particular self an eigenvector The eigenvectors of a transformation determine the directions of all invariant lines which pass through the origin A special method has been developed to solve the important equation Mv = Ay. ved. First, itis written as (M—AL)v=0. Ifthe matrix M—AL had an inverse then you could multiply by this inverse, obtaining =O. Since v #0, the conclusion must be that M—AT is singular This leads to the following method for finding eigenvalues: The eigenvectors for a matrix Mean be found by solving |M~2 1007.3 The characteri: ic equation ‘The equation [M1] =0 is called the characteristic equation for the matrix M. Example 7.3.1 Find the eigenvalues and eigenvectors of the matrix M=}-1 0 1 1201 > (3-A)[-All-A)- 2 +42 -4-6= (AA 24 SAR -k The eigenvector for the eigenvalue 2=—1 is given by 1010 Example 7.3.2 shows a case where the characteristic equation was of order 3 and had precisely three solutions. Sometimes, as in the next two examples. the characteristic equation may have fewer solutions than its order 8-. 4 > (-8-A)(12-A)+(4«25) SH 4A 40 =(4-2) = =A=2 This matrix transformation has only oneeigenvalue and associated cigenvector 102Example Find the eigenvalues and eigenvectors of the matrix M 4020-14 > -(A-3) (A+ 6)=0 > A-3or -6. For A= 3. In this case, the eigenvalue 3 has an entire plane of eigenvectors ~ any linear combination off | 0 2 and 103Show that} 7 is an eigenvector of the matrix M and find the associated cigenvalue [3 1 1] Find the characteristic equation of the matrix | 2-23 (a) Find a line of invariant points for the transformation with matrix 0305 18 8 10) (b) Hence find the eigenvalues and eigenvectors for M Find the eigenvalues and corresponding eigenvectors for the matrix M= 1 a 0 -ay The matrix Ais defined by A=) 0 4 0. whereaiis a real number ~a 0 0 (a) Given that v, =] 1] is an eigenvector of A, find the corresponding eigenvalue Ar | is another eigenvector of A, show that l ‘orresponding tov. (b) Given that eigenvalue (c) Given that the third eigenvalue Asis 6, find a corresponding eigenvectors, TAQA-NEAB, 2001] 1047.4 Diagonalisation A diagonal matrix is a square matrix which only has non-zero elements on the leading diagonal. For example. roo) foo) yl ° ° o -s! jo of ps 0 - ~ Joo The geometrical effect of @ diagonal maitix can be seen clearly. For example io of] 1] [0 0 M=|0 3 0) haseigenveetors|0|, | and 0. with associated eigenvalues |, 3 and o| 0 1 1 is mapped to 3 times itself ‘The }-axis is invariant oO} 0 © is mapped to 5 times itself. ‘The z-axis is invariant The effect of M is that of a one-way stretch, «3, in the direction of thev-axis and a one-way in the direction of the z-axis. stretch, « In fact, when a matrix transformation has a full set of eigenvectors (ic. when an 1 «17 matrix hhas srindependent eigenvectors) then the matrix is actually closely related to a diagonal matrix with the eigenvalues as its diagonal elements. This idea will first be illustrated with a simple exampleExample 7.4.1 tam=4 | 6 3) (a) Find the eigenvalues vi, v2 and associated eigenvectors Ar, 22 of M. A 0 : (b) Let We the matrix [v, _v.] and D the diagonal mais | Find VDV Solution >A=lor6. [(a)] [ Ao1 {(b)} In this case, M= DV". The proc of expressing a matriy in the form VDV. diagonalisation, where Dis a diagonal, is called An p17 matrix M which has u independent eigenvectors can be expressed as WOW", where Vis an invertible matrix and Dis a diagonal matrix V consists of the eigenvectors of M, and D consists of the eigenvalues The order of the eigenvalues in D corresponds to the order of the eigenvectors in V 106Example 7.4.2 Find an invertible matrix V and a diagonal matrix D such that Sol Using the Cheek 3, as required 121 a) When M=VDV |, the properties of M and Dare very similar. For example. viboily |=10) a M’=(VDV'\(VDV")...(VDV"} =Vb(V'V)D(VV)...(V V)DV! =VD'V In effect, the matrix Mis itself'a diagonal transformation matrix in a space where the base e. 1 [) vectors are the eigenvectors instead of the usual vectors [\ and || D If M=VDV ‘then M"=VD'V' 107Example 7.4.3 ‘The matrix M has nvectors vi and v2, with associated eigenvalues Ar and J: Find My, where vay, = bv Solution My = M(av, =A.) =a(My,)+5(Mv.) = alAy )=bl Av.) =A, (av,)~ ABV.) The idea of expressing a matrix Min the form VDV~ works in general, providing that M hat a full set of eigenvectors. This can be proved reasonably simply Let M be a 3» 3 matrix with eigenvalues Ar, 2 and 2s and associated eigenvalues v1, v2 and vs, respectively. The eigenvectors must be independent Let V be the matrix of eigenvectors [y, v. vs] My, =Ay, andsoMV=[Av Av. Ave] Because the eigenvectors are independent, the matrix Vis invertible. ‘Then V7 V= I and so 1 0] fo] Viv=)0. Vovo= 1 andV v.= 0 0) 0) 1] Then V'MV=V [Ay » Aw] “ > VOMV 0 0 A 0 SMEV0 4. 0 0 108Exercise 7C 1. Write down the eigenvalues and eigenvectors of the matrix M = VDV7, where [2 00 3020] D=/0 1 blandV= 1 0 1 jo os 11 4 | fa pv Express| inthe form VDV" where Dis a diagonal matrix 10 6) 3. The matrix Mis of the form V) 0 2 0/ Vind the determinant of M. 6 0 3) 4, Diagonalise cach of the following matrices. [Lse your answers to Questions 3, 4 and 5 of Exercise 7B] 55 32 20 2 (ay/-10 35 (b)}203 o4 0 18 8 10] 11-1 20 5 Miscellaneous exercises 7 |. A linear transformation T of 3-dimensional space is represented by the matrix. (a) Show that M has just two eigenvalues, 4, =2 and 2.= (b) Find an eigenvector corresponding to 2 (c) Show that all vector in the plane v-+)+ 2 =0 are eigenvectors corresponding to 4, (d)_ (i) Write down a vector equation of a line /. which is invariant under I (ii) State, giving a reason, whether points on Z.are invariant under T IAQANEAB, 2000] 109a . [; . | and the plane transformation T is such that b 3a 2. For constants a and h, the matrix M ¥ y ual! (a) (i) Determine the ci simplest form your answers in terms ofa and Ain their nvalues of M, givi (ii) In the case when « =3 and b= 6, show that M has no real eigenvalues. (b) A general point on the unit cirele, centre the origin, is P(cos@. sin@) (i) Show that, with a= 3 and 6 =-6, the im, form PTR cos(O-a"), Rsin(O =e) | State the values of the positive real constants Ry and 2, and find the exact value of tana. of P under T can be written in the (ii) Deseribe the locus of P” as 0 varies [AQA-NEAB. 2000], 4, (a) Find the eigenvalues and corresponding eigenvectors of M 10 (b) Hence express Min the form VDV", where Dis }0 2 0 003 5. Thematrix M=}2 41 (a) Show that M has an eigenvector | 0}. Find the corresponding eigenvalue (b) Hence find two further eigenvalues and corresponding eigenvectors. (c) Express Min the form VDV" 1109-8 6. (a) Express w,| 13) -1] inthe form VDV", where Dis a diagonal matrix 40 2 (b) Hence find M® for any integral 1 14) 7. The matrix A is given by Se a| (a) (i) Find the eigenvalues of A. (ii) Bor cach eigenvalue find a corresponding eigenvector 5 . as ~ (b) Given that U=| a write down the values of'e and h such that 3h INEAB] (a)_ (i) Show that A has only one eigenvalue. Find this eigenvalue and a corresponding eigenvector (ii) Hence, or otherwise, determine a Cartesian equation of the fixed line of T, (b) Under T, a square with area Icm? is transformed into a parallelogram with area dem?, Find the value of ¢. IAEB. 1996] 9. f inthe form VDV"', where Dis a diagonal matrix il10. The matrix Pis defined by P I (a) Show that vy; =] 1] and v are eigenvectors of Pand lind the two 0 = corresponding eigenvalues (b) Given that the third eigenvalue of P is 4, find the corresponding eigenvector vx (c) Show that v), v2 and vsare linearly independent a (d) Express the vector | h asa linear combination of vy, Vzand va with coeffictents in oc! terms of the constants, and ¢ INEAB] 11. Let Abe the matrix i (a) Determine the eigenvalues and corresponding eigenvectors of A ‘ys 10 i (b) (3) Show that A°-2A-81= 2%, where I=] 5 || and Z= D the matrix equation A°—2A-SI = Z by B, aA Bl. JAEB, 1997] (i) The matrix B= A"! By muhtiplyin or otherwise, find the values of the scalars cand f for whieh 12, The transformation T maps points (x, 3°) of the plane into image points (x,y) such that 4y+2p=14 y= 2x+7y=42 (a) (i) Find the coordinatesof the invariant point of T voy e (ii) Hence express Tin the form and Aisa 2x2 matrix (b)_ (i) Determine the eigenvalues and corresponding eigenvectors of the matr (ii) Deduce the Cartesian equations of the invariant lines of T, and prove that they are perpendicular ea full geometrical description of T JAEB. 1998] 1213. A linear transformation of 3limensional space is defined by r= Mr, where (a) Show that the transformation is singular if, and only if, (b) In the case when & = 2, show that M represents a transfomation of 3-dimensional space onto a plane, and find a Cartesian equation of this plane INEAB] 1d 14. The matrix Ais defined by A= 1k 1 11k (a) Find the determinant of A in terms of k {b) The matrix A corresponds toa linear transformation T in 3-dimensional space. When a region in 3-dimensional space is transformed by T, its volume, I” is increased by a factor of 4 to 4!” Find the possible values of k. INEAB] 15, Determine the eigenvalues and corresponding eigenvectors of the matrix A, where fp. , 6 eX x the plane transformation’T is defined by ‘ Be A (a) Write down a Cartesian equation of the line of invariant points of L (b) Show that all lines of the form y= 2x +k, where kis an arbitrary constant, are invariant lines of T. (c) Evaluate the determinant of A and explain the geometrical significance of this answer in relation to T. (d) Give @ full geometrical description of T AEB, 1998]Chapter7 Exercise 7A, o (a) MPs), ders 2 (¢) Reflection in y= xtan 2, where @ = tan 4. (a) Direction! 2) (b) a=3 ad Exercise 7B 1 Wy ul 2 1 3 i) -90 (b) 1} 90); 2, 19, 58 38 2 1 “16 4 lj-i: 619; -2) 1 0 5 = 1 5. (a) 4 (b) 1 ©) | 0 130I 1 1 1 3.6 10 0 4a) D=0 2 05 -N 0 0 0 19 2 1a 0) Fl le 0) (b)D=}0 6 0 91 00 S41 400 o2 71 (c) D=|0 1 0 10 0 00 6 ool 131(ali) 2a va = 2% (ii) a” +26 is negative wa |? 3cos0-+2sind |_| Vi3 cos(0-cr) ena=2 b)ti ang=2 "| 6 9. sino 0~9sind |" | 313 sin(0—ar) | 3 (ii) An ellipse 7 3. (a) (A WAH A+ 2) 50 =2)-1 o| 1 2) 0] {12 0 4(a) Ljt fs 2fo. 3)-1 (6) V=}1 0-1 0 L 1 ford 3 fF 1 100 3 1 1] 5. (a) 2 tb) Lf 1 (c)D=]0 2 0) V=) 1 0 1 3 [-1) oo 5 3-2 -1] -1) 0] 6. (a) V= 1, 0 2| 0 0 oj 1 0 0) jlo 0} (b) M'=VD'V D’=/0 -| 0 formodd, D’=]0 1 0) formeven 0 00) 00 oj Hence M” = M for #1 odd 5 3-1) and M2) 13-1. forsveven 26 4)7. (ai) gradient — (b)i) ‘| (c) Two-way stretch from (0,— 7) <3 in the direction of y= and »8 in the direction of y= 2x, 13. (a) [MJ=10-sk > |M|-0 >k-2 (b) 2v+2y— 14 (a) = 2k=1 (b) (K-14 k= 6 4 x | 27y-5k | 15, (a) y= Sr o * hen y kas required -Lyek | |v =2k | (c) |A]=27 Ttransforms areas by a scale factor of 27. T does not involve a reflection (d) A one-way streteh from y