Ce pe Cer | | PIR MEHAR ALI SHAH ARID AGRICULTURE UNIVERSITY RAWALPINDI Introduction to Diagonalisation of Matrics Final Project (MSC Math Fall 2015)DEDICATED We dedicate our project to our respected parents, who brought us from skies to this world and honorable teachers who took us from earth to skies. Undoubtedly, they make us enable to achieve any goal in order to get success in life. Moreover, they always have been role model for tus, We believe that success is at their feet. They have not only taught us but also inspire us.FINAL APPROVAL, This is to certify that we have read this report with full attention in which is submitted by Faiza Rafique, Awais Ahmad and Aleena Maheen. It is our judgment that this report is of sufficient standard to warrant its acceptance by Barani Institute of Sciences Sahiwal, for the degree of MSc (Math) Committee 1. Supervisor ie (Sikandar Mehmood)DECLARATION This project is Introduction to Diagonalisation of Matrices and it’s neither as @ whole nor as a part have already been developed by any person, It is further declared that we have developed this system and its interface entirely on the basis of our personal effort, made under the guidance of our project supervisor. No portion of this work presented in this report has been submitted in support of any Application for any other degree or qualification of this or any other university or institute of leaning. It is further stated that the project and all it associated documents, and records are submitted as the partial fulfillment for MSC Mathematics. We understand and transfer the copy right for this material to the Department of Methematics, Barani Institute of Sciences Sahiwal. Faiza Rafique 15-arid-4087 Awais Ahmad 15-arid-4085 Aleena Maheen 15-arid-4082 2015-2017—_— i~ Contents }, Fundamentals Page No 1.1 General introduction i 1.2 History 2 1.3 Basic definition 3 1.4 Eigenvalues and eigenvector 6 1.5 Characterization 7 2. Process of diagonalizationChapter 1 Fundamentals 1.1 General Introduction Matrix diagonalization is the process of taking a square matrix and converting it into a special type of matrix called diagonal matrix that shares the same fundamental properties of the underlying matrix. Matrix diagonalization is equivalent to transforming the underlying system of equations into a special set of coordinate axes in which the matrix takes this canonical form. Diagonalizing a matrix is equivalent to find the matrix eigenvalues, which turn out to be precisely the entries of the diagonalized matrix. Similarly, the eigenvectors make up the new set of axes corresponding to the diagonal matrix. The remarkable relationship between a diagonalized matrix, eigenvalues and eigenvectors follows from the beautiful mathematical identity the Eigen decomposition that a square matrix A can be decomposed into the very special form A=PDP™* (a) Where P is a matrix composed of the eigenvectors of A, D is the diagonal matrix constructed from the corresponding eigenvalues, and P is the matrix inverse of P. According to the i decomposition) theorem, an initial matrix equationemely powerful. The Eigen value problem is plication. For example, It'S hn models and ich as Physics, n “eigen echnique is often used and ext erable theoretical interest and wide-ranging @P! of differential equations, analyzing population growt! ponential matrix). other areas Su siderable attention © ‘or equation is engineering, where t a problem of consi crucial in solving syste calculating: of powers of Biology, Economics a vectors” their applications and t e form ms f matrices (i.e., the €x1 and Statistics have focused col heir compu In socioloBy, itations. The vect’ values” and “ele! equivalent to 2 matrix equation of th AK =b Where A is an mxn matrix, X is a column vector with n entries, and b is a column vector with m entries. an x an x2 sn} ,X=|*3], b Qn, Sm n: nk of the matrix. The number of vectors in bases for the span is now expressed as the rai 1.2 History alization goes back to ancient times! But the term “matrix” is the Latin word for womb, and it generally any place in which something is : The history of matrix diagon “Matrix” was not applied to the concept until 1850. fetains that sense in English. It can also mean more _ formed or produced. er A matrixes lie with study of systems of simultaneous linear See chit te from between 300 BC and AD 200, nine chapters of the iu Chang Suan Shu), gives the first known example of the use of matrix ations. In the treatise’s seventh chapter, “Too much and not 1's, nearly two millenary before its supposed a in 1683 or his German contemporary e invention of differential calculus, separately uses of matrix, like arrangements ofVirginia, which came to an abrupt end after an enraged Sylvester hit a newspaper-reading student with a sword stick and fled the country, believing he had killed the student. 1.3 Basic Definitions 1.3.1 Defective matrix In linear algebra, a defective matrix is a square matrix that does not have complete bases of eigenvector, and is therefore not diagonalizable. In particular, an nxn matrix is defective if and only if it does not have n linearly independent eigenvectors, A complete basis is formed by augmenting the eigenvectors with generalized eigenvectors, which are necessary for solving defective systems of ordinary differential equation and other problems. 1.3.2 Scaling (geometry) Scaling is a linear transformation enlarges (increases) or shrinks (diminishes) object factor by a scale factor that is the same in all directions. The result of uniform scaling is similar {in the geometric sense) to the original. A scale factor of | is normally allowed, so that congruent shapes are also classed as similar. Uniform scaling happens, for example, when i or reducing a photograph, or when creating a scale model of a building, car, airplane, are zero. Similarly, a square matrix is called \gonal are zero. A triangular matrix is one . A matrix that is both upper and lowerAn orthogonal matrix Qis necessarily invertible. 1.3.5Hermitian matrix Ih mathematics, @ Hermitian matrix (or self adjoint matrix) is @ square matrix with complex entries that is equal to its own conjugate transpose that is,the element in the i-th row and j-th column is equal to the complex conjugate of the element in the j+th row and i-th column, for all the indices i and j, ay Or A=Ae Hermitian matrices can be understood as the com; the conjugate transpose of a matrix A is denote: written concisely as, plex extension of real symmetric matrices. If dd byA‘, then the Hermitian property can be A=Ay Hermitian matrices are named after Charles Hermite, who demonstrated in 1885 that matrices of this form share a property with real symmetric mati trices of always having real eigenvalues,is nilpotent, with linear combination of 23,%2,%3)-%n onalizable, then the system can be un-coupled in that each %The matrix is denoted by diag(9,8,6) 1.3.10 Diagonalizable matrix The matri A is diagonalizable if it is similar to a diagonal matrix, in other words: If there Is a diagonal matrix D and an invertible matrix P such that P~1AP = D. If V isa finite dimensional vector space, then a linear mapping T: VV is called diagonalizable if there exists an ordered bases of V with respect to which T is represented by a diagonal matrix. Diagonalization is the process of finding a corresponding diagonal matrix for a diagonalizable matrix or linear mapping. A square matrix that is not diagonalizable is called defective. Diagonalizable matrices and maps are of interest because diagonal matrices are especially easy tohandle: their eigenvalues and eigenvectors are known and one can raise a diagonal matrix to power by simply raising the diagonal entries to that same power, 1.4 Eigenvalues and eigenvectors Suppose that A is a square matrix. if for some non-zero vector X, AX = AX. A\ called an eigenvector for eigenvalue A eigenvalue a. The number A is said to be an eigenvalue of matrix A ‘ny non-zero vector X for which this equation holds is or an eigenvector of matrix A Corresponding to 1.4.1 Process of finding eigenvalues and eigenvectors termine whether A is an eigenvalue of A, we need to determine whether there solution to the matrix equation AX = 2 X. Note that the matrix equation the standard form, since the right-hand side is not a fixed vector b, but vever, we can rewrite it in standard form. Note that AX =AIX, where t quivalent to AX = AIX, or AX AIX = 0, which0 a i and the characteristic polynomial is la- a=? ae | MRED =(-a)@-a) =? -3A4+2-2 = 4-32 0 the eigenvalues are the solutions of 2” — 34 = 0. To solve this, simply observe that the equation is A(A — 3) = 0 with solutions 2 = 0 and A = 3, Hence the eigenvalues of A are 0 and a To find an eigenvector for the eigenvalue 2, we have to find a solution to (A ~ l)x = 0, other than the zero vector. This is easy, since for a particular value of 2 , all we need to do is solve a simple linear system we illustrate by finding the eigenvectors for the matrix of the example just given. ab a the eigenvalues are 0 and 3. To find an eigenvector for eigenvalue 0 we A Ol)x =0: that is, Ax =0, or babelae nn __ ee a EERE of eigenvectors of A. if such base has been found, one can form the matrix P having these bases vectors as columns and P~*AP will be diagonal matrix. The diagonal entries of this matrix are the eigenvalues of A. A linear map T: VV is diagonalizable if and only if the sum of the dimensions of its eigenspaces is equal to dim(V), which is the case if and only if there exist bases of V consisting of eigenvectors of T, with respect to such a bases T will be represented by a diagonal matrix. The diagonal entries of this matrix are the eigenvalues of T. Another characterization: A matrix or linear map is diagonalizable over the field F if and only if its minimal polynomial is the product of distinct linear factors over F, (Put in another way, a matrix is diagonalizable if and only if all of its elementary divisors are linear).The following sufficient (but not necessary) condition is often useful. A matrix of order nxn is diagonalizable over a field Fifithas n distinct eigenvalues in F i.e, if its characteristic polynomial has n distinct roots in F; however, the converse may be false. A linear map T: V-> V with n=dim(V) is diagonalizable if it has n distinct eigenvalues i.e. characteristic polynomial has n distinct roots in F.— ts A ——TOOy—————— Chapter # 2 process of Diagonalisation 2.1 Diagonalisation Ifa matrix A is diagonalizable that is P>AP = ‘Then Dao Rate) “ ra i) 2) where (i= 1,2,3,...,.n) of P are right eigenvectors of A, and the corresponding diagonal entries eigenvalues. The invariability of P also suggests that the eigenvectors are form bases of F. This is the necessary and sufficient condition for oact ch of ioe ation. The Tow vectos of PA set of matrices are said to be simultaneously diagonalizable if there exists 2 single ' invertible matrix P such that P~*AP is a diagonal matrix for every A in the set.The following theorem characterizes simultaneously diagonalizable matrices: A set of diagonalizable matrices commutes if and only if the set is simultaneously diagonalizable. The set of all nxn diagonalizable (over C) with n > 1 is not simultaneously diagonalizable. For instance, the matrices Opera fo olf ol are diagonalizable but not simultaneously diagonalizable because they do not commute. A set consists of commuting normal matrices if and only if it is simultaneously diagonalizable by 2 unitary matrix ie., there exist a unitary matrix U such that U-?AU is diagonal for every A in the set. 2.3 Diagonalizable matrices Involution are diagonalizable over the real (and indeed any field of characteristic not 2), with #1 on the diagonal. Finite order endomorphism are diagonalizable over C (or any algebraically closed field where the characteristics of the field does not divide the order of the endomorphism) with roots of unity on the diagonal. This follows since the minimal polynomial is separable, because the roots of unity are distinct. Projection is diagonalizable, with O's and V's on the diagonal. Real symmetric matrices are diagonalizable by orthogonal matrices L.e., given a real symmetric matrices A, Q'AQ is the diagonal for some orthogonal matrix Q. More generally, matrices are iagonalizable by unitary matrices if and only if they are normal. In the case of real symmetric ee that A = At so clearly AA‘ = A'A holds. Examples of normal matrices ere the skew-symmetric) matrices and Hermition matrices (or skew-Hermitionas. kl LS Lae ccuauaannnmmmmmmmmmememmmmemmme amma | Aisa 3X 3 matrix with 3 different eigenvalues; therefore, it is diagonalizable. Note that if there are exactly n distinct eigenvalues in an nxn matrix then this matrix is diagonalizable. These eigenvalues are values that will appear in the diagonalised form of matrix A, so by finding the eigenvalues of A we have diagonalized it. We could stop here, but it is a good check to use the eigenvectors to diagonlize A. «Feel Avy = Av, The eigenvectors of A are One can easily check that Now, let P be the matrix with these eigenvectors as its columns, -1 0 -1 = 0 2 iee 2 Note there is no preferred order of the eigenvectors in P, changing the order of the “igenvectors in P just changes the order of the eigenvalues in the diagonalized form of A,[3] ‘Then A is diagonalizable, as a simple computation confirms, having calculated P~* using any Suitable method. OS OP Bo ait © a OG 22 alle 3 O}|-1 0 O}=jo 2 0 1 ollz -4 2 Al Thy) @But if we set, 01 a 0 lo olen) -1 0 o -1 Then, We have seen that if A and B are similar, then A” can be expressed easily in term of B. indeed, if we have A=P~*BP, then we have A” =P-*B” P. In particular, if D is a diagonal matrix, D® is easy to evaluate. This is an application of diagonalization. In fact, above procedure may be used tofind the square root and cubic root of a matrix. Indeed, consider the matrix above 1-1 1 A (0 —2 1 0 0 =12 Set Then Hence A=ppp-1 Set,Example 2 The matrix A has eigen values 1,2, and -1 with corresponding eigenvector (1,1,0), (1,2,1) and 0,1,2). Find A and compute AS. Solution:- Since A has three distinct eigenvalues, it is diagonalizable and thus A = TDT~*, where D is a diagonal matrix having the eigenvalues of A on the main diagonal and T has eigenvectors of A its columns. Ono loo poe Ei " Pa Sea Reo Ses So We compute T-* 110100 110100 (121010)~ (011110) 012001 012001 Subtract the 1" row from the 2 row &Subtract 2" row from 3? row 110100 110100 011110)~(010221 oo01111 oo01111 & Subtract the 3"¢ row from 24 row &subtract the 2" row from the 1st row,A=Tp5r 110\/100)\/3 -2 1 -(121 32 0)(-2 a) 012/\00-1/\1 -1 1 110\/3 -2 14 (i 2 :)(—<« 64 =) O12/\-1 1 -1 —61 62 -31 -(-i26 127 64 66 66 —34, 25 Non diagonalizable matrices In general, a rotation matrix is not diagonalizable over the real, but all rotation ‘matrices are diagonalizable over the complex field. Even if a matrix not diagonalizable, it is always possible to “do the best one can’, and find a matrix with the same properties consisting of eigenvalues on the leading diagonal, and either ones or zeroes on the super diagonal (known as Jordan normal form). Some matrices are not diagonalizable over any field, most notably nonzero nilpotent ‘Tatrices. This happens more generally if the algebraic and geometric multiplicities of an “genvalue do not coincide, For instance, consider ee cl ol diagonalizable there is no matrix U such that U-1CU is a diagonal matrix Cigenvalue (namely zero) and this eigenvalue has algebraic multiplicity 2 uultiplicity 1.Some real matrices are not diagonalizable over the real. ince the matrix, This matrix is not Indeed, ¢ has one and geometric m Consider for instayCee: -1 0. the matrix B does not have any real eigenvalues, so there is no real matrix Q such thatQ2BQ is a diagonal. However, we can diagonalise B if we allow complex numbers Indeed, ifwe take then Q~'BQ is diagonal. Note that the above example show that the sum of diagonalizable matrices need not be diagonalizable. 2.6 Applications of diagonalization Diagonalization can be used to compute the power of a matrix A efficiently, provided the matrix is diagonalizable. Suppose we have found that PAP =D 1s diagonal matrix. Then, as the matrix product is associative, Abe (PDP) E = (PDP™). (PDP) ... (PDP~*) = PD(P™*P)D(P~1P) .... (P-*P)DP* =PDkP™* And the latter is easy to calculate since it involves the power of diagonal matrix. This approach ‘an be generalized to matrix exponential other matrix functions since they can be defined as Power series. This is particularly useful in finding closed form expressions for terms of linear ‘cursive sequences, 26.1 Particular application Fi "example, consider the following matrix la b-a m=[o°;' Caleutatix ; Ne the various powers of M reveals a surprising patternlee |. ms =[2 eat Me iB by as The above phenomenon can be explained by diagonalizing M. To accomplish this, we need a basis of R consisting of eigenvectors of M. One such eigenvector basis is given by w=[Ehe. o= [pen Where e,denotes the standard basis of”. The reverse change of basis is given by e=u , e2=v-u straight forwarded calculations show that Mu=au , Mv=av Thus, a and b are the eigenvalues corresponding to u and v respectively. By linearity of matrix multiplication, we have M"a=a"u,M"v=b"v Switching back to the standard basis, we have Me, = M"u=a"e, Me, = M"(v —u) = b"v —a"u = (b" — a")e, + b"e The preceding relations, expressed in matrix form are fe eae | There by explaining above phenomenon. Application of Diagonalization of the Coefficient Matrices to Differential Equations Consider first order differential equations Yi=Yi-Y2#2Y3, Y2=3Y/+4Y3, Y3=2Y1+Y2 We can wite these equations as wbl = a1 ie wrfs 0 ' Zee The eigenvalues of [A] are 1.303,-2.303 and 2. The corresponding eigenvectors are: 0.172) (—0.557] | 0 0.891 |, )—0.467], |0.894| 042 J 10.687 | loz, ‘The model transformation matrix is [m] - [est —0.467 0.894 042 0.687 0,447. __ [Mi Diagonalise [A] and setingly]={M[Z}transform [y]-[A][y1 into — 2] = ullz) -0.172-0.557 023 = ce? Therefore, ae 303 (2) =|be-2303¢ raze sincelY) = (M1 £21 [YF] 0.891 —0.467 0.894] |be~2303¢ 0.172-0.557 0 ee] 0.42 0.687 0.447] ce In terms of component, the general solution is yy = -0.172a€%49 -0,557be-2 303 Ya = 0.8914e13* -0.467be-2293"+.0,894ce% Ya = OA2ae12°2* -0,687be-23033462 Quantum mechanical application In quantum mechanical and quantum chemical cor ‘one of the most frequently applied numerical ‘A very common approximation is which the Schrodinger equation can be formulaConclusion We have concluded that solution by diagonalization will always work provided. We can find in nearly independent eigenvectors ofthe msn matrix be the eigenvalue of A could an istnet, complex or repeated. But the method fails when A has repeated ei a linearly independent eigenvalue cannot be found. Of course, in the last iagonalizable. So in such conditions, one cannot use diagonalization lution of matrices. Diagonalization can also be used to so h