next up previous
Next: dd_sym Up: dd_sym Previous: dd_sym

3. An application

A quadratic expression such as $ 6 x ^ 2 + 4 x y + 9y ^2$ is said to be a ``quadratic form''.

A quadratic form can be rewritten as x$ ^ t A$   x, where $ A$ is symmetric and x$ = \left[\begin{array}{c}x\\  y\end{array}\right]$ (a column vector) so x$ ^ t$ is a row vector. For example, $ 6x ^ 2 + 4 xy + 9y ^ 2 =
\left[\begin{array}{cc}x&y\end{array}\right] \left[\b...
...{rr}6&2\\  2&9\end{array}\right] \left[\begin{array}{c}x\\  y\end{array}\right]$. (Try multiplying out to check.) The coefficient of $ xy$ is split in half and each half becomes an off-diagonal entry.

To analyze the graph of an equation such as $ 6x ^ 2 + 4xy + 9y ^ 2 = 1$, write it in matrix form, diagonalize the matrix, and make the substitution x$ = R$r, where $ R$ is a rotation matrix.

Example: Describe the graph of $ 3 x ^ 2 + 2 x y + 3 y ^ 2 = 1$.

Solution: Rewrite as x$ ^ t A$   x$ = 1$ for $ A = \left[\begin{array}{cc}3&1\\  1&3\end{array}\right]$. We saw that $ R ^ {-1} A R = D$ for $ R = R _ {45 ^ \circ}$ and $ D = \left[\begin{array}{rr}4&0\\  0&2\end{array}\right]$. Substitute in x$ = R$   r. We get $ (R$r$ ) ^ t A (R$r$ ) = 1$, or r$ ^ t (R ^ t A R)$   r$ = 1$. Since $ R ^ t = R ^ {-1}$, we get r$ ^ t D$   r$ = 1$, or $ 4 r ^ 2 + 2 s ^ 2 = 1$. This can be rewritten as

$ {\frac{\displaystyle r ^ 2}{\displaystyle a ^ 2}} + {\frac{\displaystyle s ^ 2}{\displaystyle b ^ 2}} = 1$ for $ a = {\frac 1 2}$, $ b = {\frac 1 2}\sqrt 2$,

so the shape is an ellipse with semimajor axis $ {\frac 1 2}\sqrt 2$ and semiminor axis $ {\frac 1 2}$. But the ellipse is slanted. How? Since $ a < b$, the end of the semimajor axis is where $ r =
0$, $ s = b = {\frac 1 2}\sqrt 2$; here x$ = R \left[\begin{array}{c}0\\  {\frac 1 2}\sqrt
2\end{array}\right] = [$v$ _ 1 \vert$   v$ _ 2] \left[\begin{array}{c}0\\  {\frac 1 2}\sqrt 2\end{array}\right] = {\frac 1 2}\sqrt 2$   v$ _ 2$, where v$ _ 2$ is the second eigenvector, which is the same as the second column of $ R$. Therefore the end of the semimajor axis of $ R$ is at $ {\frac 1 2}\sqrt 2 \left[\begin{array}{c}-{\frac 1 2}\sqrt
2\\  {\frac 1 2}\sq...
...y}\right] = \left[\begin{array}{c}-{\frac 1 2}\\  {\frac 1 2}\end{array}\right]$. You can check that this point is on the graph.



Problem DD-7. In this example, find the end of the semiminor axis.



Problem DD-8. (a) In general, the transformed equation is $ \lambda _ 1 r ^ 2
+ \lambda _ 2 s ^ 2 = 1$. Explain why, by discussing the substitution.

(b) How can you tell from the eigenvalues whether the equation is going to be an ellipse or a hyperbola or maybe some degenerate case? (In fact, what kinds of graphs can arise from (a)?)



Problem DD-9. (a) Describe the shape of the graph of $ 6x ^ 2 + 4xy + 9y ^ 2 = 1$. If it is an ellipse, give semimajor and semiminor axis lengths. (For this you can use eigenvalues alone.)

(b) How is this graph slanted? (You'll need the eigenvectors.)




next up previous
Next: dd_sym Up: dd_sym Previous: dd_sym
Kirby A. Baker 2001-11-26