Consider the linear eigenproblem, 𝐴𝑥=𝜆𝑥, for the matrix
D=[1 1 2
2 1 1
1 1 3 ]
1. Solve for the largest eigenvalue by the direct method using the secant method. Let 𝜆^((0))=5 and 𝜆^((1))=4
2. Solve for the eigenvalues by the QR method
1.
"f(\\lambda_0)=f(5)=\\begin{vmatrix}\n 1-5 & 1&1 \\\\\n 2 & 1-5&1\\\\\n1&1&3-5\n\\end{vmatrix}=-4\\cdot7+5-2=-25"
"f(\\lambda_1)=f(4)=\\begin{vmatrix}\n 1-4 & 1&1 \\\\\n 2 & 1-4&1\\\\\n1&1&3-4\n\\end{vmatrix}=-3\\cdot2+3+5=2"
slope = "\\frac{f(\\lambda_1)-f(\\lambda_0)}{\\lambda_1-\\lambda_0}=\\frac{2+25}{4-5}=-27"
the largest eigenvalue:
"\\lambda_2=\\lambda_1-\\frac{f(\\lambda_1)}{slope}=4+2\/27=4.074"
2.
"d_1=\\begin{pmatrix}\n 1 \\\\\n 2 \\\\\n1\n\\end{pmatrix}" , "d_2=\\begin{pmatrix}\n 1 \\\\\n 1\\\\\n1\n\\end{pmatrix}" , "d_3=\\begin{pmatrix}\n 2 \\\\\n 1 \\\\\n3\n\\end{pmatrix}"
"||d_1||=\\sqrt{1+2^2+1}=2.449"
"q_1=d_1\/||d_1||"
"q_1^T=\\begin{pmatrix}\n 0.408 & 0.816&0.408 \\\\\n \n\\end{pmatrix}"
"d'_2=d_2-(q_1^Td_2)q_1"
"d'_2=\\begin{pmatrix}\n 1 \\\\\n 1\\\\\n1\n\\end{pmatrix}-\\begin{pmatrix}\n 0.408 & 0.816&0.408 \\\\\n \\end{pmatrix}\\begin{pmatrix}\n 1 \\\\\n 1\\\\\n1\n\\end{pmatrix}\\begin{pmatrix}\n 0.408 \\\\\n 0.816\\\\\n0.408\n\\end{pmatrix}=\\begin{pmatrix}\n 0.834 \\\\\n 0.334\\\\\n0.834\n\\end{pmatrix}"
"||d'_2||=1.226"
"q_2^T=d'_2\/||d'_2||=\\begin{pmatrix}\n 0.680 & 0.272&0.680 \\\\\n \n\\end{pmatrix}"
"d'_3=d_3-(q_1^Td_3)q_1-(q_2^Td_3)q_2=\\begin{pmatrix}\n 2 \\\\\n 1\\\\\n3\n\\end{pmatrix}-\\begin{pmatrix}\n 0.408 & 0.816&0.408 \\\\\n \\end{pmatrix}\\begin{pmatrix}\n 2 \\\\\n 1\\\\\n3\n\\end{pmatrix}\\begin{pmatrix}\n 0.408 \\\\\n 0.816\\\\\n0.408\n\\end{pmatrix}--\\begin{pmatrix}\n 0.680 & 0.272&0.680 \\\\\n \\end{pmatrix}\\begin{pmatrix}\n 2 \\\\\n 1\\\\\n3\n\\end{pmatrix}\\begin{pmatrix}\n 0.680 \\\\\n 0.272\\\\\n0.680\n\\end{pmatrix}=\\begin{pmatrix}\n 2- 0.333-0.924\\\\\n 1-0.666-0.074\\\\\n3-0.499-1.387\n\\end{pmatrix}=\\begin{pmatrix}\n 0.743\\\\\n 0.260\\\\\n1.114\n\\end{pmatrix}"
"||d'_3||=1.364"
"q_3=d'_3\/||d'_3||"
"q_3^T=\\begin{pmatrix}\n 0.545 & 0.191&0.817 \\\\\n \n\\end{pmatrix}"
"Q^{(0)}=\\begin{pmatrix}\n q_1 & q_2&q_3 \\\\\n \n\\end{pmatrix}=\\begin{pmatrix}\n 0.408 & 0.680&0.545 \\\\\n 0.816 & 0.272&0.191\\\\\n0.408&0.680&0.817\n\\end{pmatrix}"
for "R^{(0)}" :
"r_{11}=||d'_1||=2.449,r_{22}=||d'_2||=1.226,r_{33}=||d'_3||=1.364"
"r_{12}=q_1^Td_2=1.632,r_{13}=q_1^Td_3=2.856,r_{23}=q_2^Td_3=3.672"
"R^{(0)}=\\begin{pmatrix}\n 2.449 & 1.632&2.856 \\\\\n 0 & 1.226&3.672\\\\\n0&0&1.364\n\\end{pmatrix}"
"D^{(1)}=R^{(0)}Q^{(0)}=\\begin{pmatrix}\n 2.449 & 1.632&2.856 \\\\\n 0 & 1.226&3.672\\\\\n0&0&1.364\n\\end{pmatrix}\\begin{pmatrix}\n 0.408 & 0.680&0.545 \\\\\n 0.816 & 0.272&0.191\\\\\n0.408&0.680&0.817\n\\end{pmatrix}="
"=\\begin{pmatrix}\n 3.496 & 4.051&4.000 \\\\\n 2.499 & 2.830&3.234\\\\\n0.557&0.928&1.114\n\\end{pmatrix}"
The diagonal elements in matrix "D^{(1)}" are the first approximation to the eigenvalues of matrix "D".
So,
"\\lambda_1=3.496,\\lambda_2=2.830,\\lambda_3=1.114"
Comments
Leave a comment