Ode 11
Ode 11
Ode 11
Ghorai
Lecture XI
Euler-Cauchy Equation
If the ODE is of the form where a, b and c are constants; then (1) is called homogeneous Euler-Cauchy equation. Two linearly independent solutions (i.e. basis) depend on the quadratic equation am2 + (b a)m + c = 0. (2)
Equation (2) is called characteristic equation for (1). The ODE (1) is singular at x = 0. Hence, we solve (1) for x = 0. We consider the case when x > 0. Theorem 1. (i) If the roots of (2) are real and distinct, say m1 and m2 , then two linearly independent (LI) solutions of (1) are xm1 and xm2 . Thus, the general solution to (1) is y = C1 xm1 + C2 xm2 . (ii) If the roots of (2) are real and equal, say m1 = m2 = m, then two LI solutions of (1) are xm and xm ln x. Thus, the general solution to (1) is y = (C1 + C2 ln x)xm . (iii) If the roots of (2) are complex conjugate, say m1 = + i and m2 = i, then two real LI solutions of (1) are x cos( ln x) and x sin( ln x). Thus, the general solution to (1) is y = x C1 cos( ln x) + C2 sin( ln x) . Proof: We have seen that the trial solution for a constant coecient equation is emx . Now since power of xm is reduced by 1 by a dierentiation, let us take xm as trial solution for (1). For convenience, (1) is written in the operator form L(y) = 0, where d d2 + bx + c. L ax 2 dx dx
2
We also sometimes write L as L ax2 D2 + bxD + c, where D = d/dx. Now L(xm ) = am(m 1) + bm + c xm = p(m)xm , (3)
where p(m) = am2 + (b a)m + c. Thus, xm is a solution of (1) if p(m) = 0. (i) If p(m) = 0 has two distinct real roots m1 , m2 , then both xm1 and xm2 are solutions of (1). Since, m1 = m2 , they are also LI. Thus, the general solution to (1) is y = C1 xm1 + C2 xm2 .
Solution: The characteristic equation is m2 2m 3 = 0 m = 1, 3. The general solution is y = C1 /x + C2 x3 (ii) If p(m) = 0 has real equal roots m1 = m2 = m, then xm is a solution of (1). To nd the other solution, note that if m is repeated root, then p(m) = p (m) = 0. This suggests dierentiating (3) w.r.t. m. Since L consists of dierentiation w.r.t. x only, m x = L(xm ln x). L(xm ) = L m m Now L(xm ln x) = p (m) + p(m) ln x xm , where represents the derivative. Since, m is a repeated root, the RHS is zero. Thus, xm ln x is also a solution to (1) and it is independent of xm . Hence, the general solution to (1) is y = (C1 + C2 ln x)xm . Example 2. Solve x2 y 3xy + 4y = 0 Solution: The characteristic equation is m2 4m + 4 = 0 m = 2, 2. The general solution is y = (C1 + C2 ln x)x2 . (iii) If p(m) = 0 has complex conjugate roots, say m1 = + i and m2 = i, then two LI solutions are Y1 = x(+i) = x ei ln x , and Y2 = x ei ln x .
But these are complex valued. Note that if Y1 , Y2 are LI, then so are y1 = (Y1 + Y2 )/2 and y2 = (Y1 Y2 )/2i. Hence, two real LI solutions of (1) are y1 = x cos( ln x) and y2 = x sin( ln x). Thus, the general solution to (1) is y = x C1 cos( ln x) + C2 sin( ln x) . Example 3. Solve x2 y 3xy + 5y = 0 Solution: The characteristic equation is m2 4m + 5 = 0 m = 2 i. The general solution is y = x2 C1 cos(ln x) + C2 sin(ln x) Comment 1: The solution for x < 0 can be obtained from that of x > 0 by replacing x by x everywhere. Comment 2: Homogeneous Euler-Cauchy equation can be transformed to linear constant coecient homogeneous equation by changing the independent variable to t = ln x for x > 0. Comment 3: This can be generalized to equations of the form a(x + )2 y + b(x + )y + cy = 0. In this case we consider (x + )m as the trial solution.
S. Ghorai
where a, b and c are constants; then (4) is called nonhomogeneous Euler-Cauchy equation. We can use the method of variation of parameters as follows. First divide (4) by ax2 so that the coecient of y becomes unity: y + c b y + 2 y = r(x), ax ax (5)
where r(x) = r(x)/ax2 . Now we already know two LI solutions y1 , y2 of the homoge neous part. Hence, the particular solution to (4) is given by yp (x) = y1 (x) y2 (x)r(x) dx + y2 (x) W (y1 , y2 ) y1 (x)r(x) dx. W (y1 , y2 )
Thus, the general solution to (4) is y(x) = C1 y1 (x) + C2 y2 (x) + yp (x). Example 4. Comment: In few cases, it can be solved also using method of undetermined coecients. For this, we rst convert it to constant coecient liner ODE by t = ln x. If the the transformed RHS is of special form then the method of undetermined coecients is applicable. Example 5. Consider x2 y xy 3y = ln x , x x > 0.
The characteristic equation is m2 2m 3 = 0 m = 1, 3. Hence y1 = 1/x and y2 = x3 . Hence, yp (x) = y1 (x)u(x) + y2 (x)v(x) where u(x) = y2 (x)r(x) dx, W (y1 , y2 ) v(x) = y1 (x)r(x) dx. W (y1 , y2 )
Now W (y1 , y2 ) = 4x and r(x) = ln x/x3 ! Hence, u(x) = v(x) = ln x (ln x)2 dx = 4x 8
ln x ln x 1 dx = 5 4 4x 16x 64x4
S. Ghorai Hence, 1 (ln x)2 ln x 8x 16x 64x Hence the general solution is y = c1 y1 + c2 y2 + yp , i.e. yp (x) = y(x) = A (ln x)2 ln x + Bx3 . x 8x 16x
Note that last term of yp is absorbed with y1 . Aliter: Let us make the transformation t = ln x. Then the given transformed to y 2y 3y = tet , where = d/dt. This is the same problem we have solved in lecture 9 using method of undetermined coecients. The solution is (see lecture 9) y(t) = C1 et + C2 e3t which in terms of original x variable becomes y(x) = C1 ln x + C 2 x3 (2 ln x + 1), x 16x tet (2t + 1), 16