Linear ODEs and Special Functions
CSIR NET Physical Sciences: first- and second-order linear ODEs, Frobenius series, and Legendre, Hermite, Laguerre and Bessel functions with practice.
- 18 min
On this page · 13 sections
Prerequisites
- Linear Algebra, Matrices and Eigenvalues: linear independence, and systems of ODEs as eigenvalue problems.
- The Laplacian in spherical and cylindrical coordinates, from Dimensional Analysis and Vector Calculus.
- Existence, uniqueness and variation of parameters at proof level: Ordinary Differential Equations in Mathematical Sciences.
The Hook: One Equation, Four Families
First-Order Linear ODEs
The standard form is
$$y' + P(x)\,y = Q(x).$$Multiply by the integrating factor \( \mu(x) = e^{\int P\,dx} \); the left side becomes \( (\mu y)' \), so
$$\boxed{y = \frac{1}{\mu(x)}\left[\int\mu(x)Q(x)\,dx + C\right]}$$Example. \( y' + y/x = x^2 \): \( \mu = x \), \( (xy)' = x^3 \), \( y = x^3/4 + C/x \).
Physics example. An \( LR \) circuit switched onto a battery \( V \): \( L\,di/dt + Ri = V \), \( \mu = e^{Rt/L} \), and with \( i(0) = 0 \), \( i = (V/R)\left(1 - e^{-Rt/L}\right) \). The time constant \( L/R \) is the only time scale, as dimensional analysis would predict.
Two related forms reduce to this one:
- Bernoulli \( y' + Py = Qy^n \): put \( v = y^{1-n} \), giving \( v' + (1 - n)Pv = (1 - n)Q \), which is linear.
- Exact \( M\,dx + N\,dy = 0 \) with \( \partial_yM = \partial_xN \): there is \( F(x, y) \) with \( F_x = M \), \( F_y = N \), and the solution is \( F = C \).
Second-Order Linear ODEs
$$y'' + P(x)y' + Q(x)y = R(x)$$The general solution is \( y = c_1y_1 + c_2y_2 + y_p \): two independent solutions of the homogeneous equation plus one particular solution.
The Wronskian
$$W(y_1, y_2) = y_1y_2' - y_2y_1'$$If \( W \ne 0 \) at one point of an interval where \( P, Q \) are continuous, the solutions are independent. Abel’s formula gives \( W \) without knowing the solutions:
$$\boxed{W(x) = C\exp\left(-\int P(x)\,dx\right)}$$So \( W \) is either never zero or identically zero on the interval.
Constant coefficients
For \( ay'' + by' + cy = 0 \), try \( y = e^{mx} \): \( am^2 + bm + c = 0 \).
| Roots | Solution |
|---|---|
| Real, distinct \( m_1, m_2 \) | \( c_1e^{m_1x} + c_2e^{m_2x} \) |
| Real, repeated \( m \) | \( (c_1 + c_2x)e^{mx} \) |
| Complex \( \alpha \pm i\beta \) | \( e^{\alpha x}(c_1\cos\beta x + c_2\sin\beta x) \) |
The damped oscillator
\( m\ddot x + b\dot x + kx = 0 \), or \( \ddot x + 2\gamma\dot x + \omega_0^2x = 0 \) with \( \gamma = b/2m \), \( \omega_0 = \sqrt{k/m} \). The roots are \( -\gamma \pm \sqrt{\gamma^2 - \omega_0^2} \).
| Regime | Condition | Motion |
|---|---|---|
| Underdamped | \( \gamma < \omega_0 \) | \( e^{-\gamma t}\cos(\omega_dt + \phi) \), \( \omega_d = \sqrt{\omega_0^2 - \gamma^2} \) |
| Critically damped | \( \gamma = \omega_0 \) (\( b = 2\sqrt{mk} \)) | \( (A + Bt)e^{-\gamma t} \); fastest return without overshoot |
| Overdamped | \( \gamma > \omega_0 \) | sum of two decaying exponentials |
The quality factor of a lightly damped oscillator is \( Q = \omega_0/2\gamma \).
Particular solutions
Undetermined coefficients works when \( R(x) \) is a polynomial, exponential, sine/cosine or a product of these: guess the same form. If the guess already solves the homogeneous equation (resonance), multiply it by \( x \) (or \( x^2 \) for a double root).
Example. \( y'' + 4y = \sin2x \). The homogeneous solutions are \( \cos2x \), \( \sin2x \), so try \( y_p = x(A\cos2x + B\sin2x) \). Substituting gives \( -4A\sin2x + 4B\cos2x = \sin2x \), so \( A = -\tfrac{1}{4} \), \( B = 0 \), and \( y_p = -\tfrac{x}{4}\cos2x \): an amplitude that grows linearly, the signature of resonance.
Variation of parameters works for any \( R \):
$$y_p = -y_1\int\frac{y_2R}{W}\,dx + y_2\int\frac{y_1R}{W}\,dx.$$Euler–Cauchy (equidimensional) equations
\( x^2y'' + axy' + by = 0 \). Try \( y = x^m \): \( m(m - 1) + am + b = 0 \). Repeated root \( m \) gives \( x^m \) and \( x^m\ln x \); complex roots \( \alpha \pm i\beta \) give \( x^\alpha\cos(\beta\ln x) \) and \( x^\alpha\sin(\beta\ln x) \). The substitution \( x = e^t \) turns the equation into one with constant coefficients. The radial part of Laplace’s equation in spherical coordinates, \( r^2R'' + 2rR' - l(l+1)R = 0 \), is of this type, with solutions \( r^l \) and \( r^{-(l+1)} \).
Series Solutions and the Frobenius Method
For \( y'' + P(x)y' + Q(x)y = 0 \), the point \( x_0 \) is
- ordinary if \( P \) and \( Q \) are analytic there: two independent power-series solutions exist, converging at least up to the nearest singularity of \( P \) or \( Q \) in the complex plane;
- regular singular if \( (x - x_0)P \) and \( (x - x_0)^2Q \) are analytic: try a Frobenius series \( y = \sum_k a_k(x - x_0)^{k+r} \);
- irregular singular otherwise.
Substituting the Frobenius series at \( x_0 = 0 \), the lowest power gives the indicial equation
$$r(r - 1) + p_0r + q_0 = 0, \qquad p_0 = \lim_{x\to0}xP, \quad q_0 = \lim_{x\to0}x^2Q.$$If the roots \( r_1 \ge r_2 \) do not differ by an integer, both give solutions. If they differ by an integer (including zero), the larger always gives one; the second may contain a \( \ln x \) term. For Bessel’s equation the indicial roots are \( \pm\nu \), which is why \( J_\nu \) and \( J_{-\nu} \) are independent only for non-integer \( \nu \).
Special Functions
All four families are eigenfunctions of Sturm–Liouville problems, so each is orthogonal with a weight on its natural interval. That one fact explains half of what is tested.
| Family | Equation | Interval | Weight |
|---|---|---|---|
| Legendre \( P_n \) | \( (1 - x^2)y'' - 2xy' + n(n+1)y = 0 \) | \( [-1, 1] \) | 1 |
| Hermite \( H_n \) | \( y'' - 2xy' + 2ny = 0 \) | \( (-\infty, \infty) \) | \( e^{-x^2} \) |
| Laguerre \( L_n \) | \( xy'' + (1 - x)y' + ny = 0 \) | \( [0, \infty) \) | \( e^{-x} \) |
| Bessel \( J_\nu \) | \( x^2y'' + xy' + (x^2 - \nu^2)y = 0 \) | \( [0, a] \) | \( x \) |
Legendre polynomials
First few: \( P_0 = 1 \), \( P_1 = x \), \( P_2 = \tfrac{1}{2}(3x^2 - 1) \), \( P_3 = \tfrac{1}{2}(5x^3 - 3x) \), \( P_4 = \tfrac{1}{8}(35x^4 - 30x^2 + 3) \).
Rodrigues formula: \( P_n(x) = \dfrac{1}{2^nn!}\dfrac{d^n}{dx^n}(x^2 - 1)^n \).
Generating function:
$$\boxed{\frac{1}{\sqrt{1 - 2xt + t^2}} = \sum_{n=0}^{\infty}P_n(x)\,t^n, \qquad \lvert t\rvert < 1}$$This is why Legendre polynomials appear in the multipole expansion: \( \dfrac{1}{\lvert\mathbf r - \mathbf r'\rvert} = \sum_l\dfrac{r_<^l}{r_>^{l+1}}P_l(\cos\gamma) \).
Orthogonality:
$$\boxed{\int_{-1}^{1}P_m(x)P_n(x)\,dx = \frac{2}{2n + 1}\,\delta_{mn}}$$Recurrences: \( (n + 1)P_{n+1} = (2n + 1)xP_n - nP_{n-1} \) (Bonnet), and \( P_{n+1}' - P_{n-1}' = (2n + 1)P_n \).
Special values: \( P_n(1) = 1 \), \( P_n(-1) = (-1)^n \), \( P_n(-x) = (-1)^nP_n(x) \); \( P_n(0) = 0 \) for odd \( n \), with \( P_2(0) = -\tfrac{1}{2} \), \( P_4(0) = \tfrac{3}{8} \).
Expanding a polynomial. Since \( P_n \) has degree \( n \), any polynomial of degree \( N \) is a finite combination of \( P_0, \ldots, P_N \). Work from the top degree down:
$$x^2 = \tfrac{1}{3}P_0 + \tfrac{2}{3}P_2, \qquad x^3 = \tfrac{3}{5}P_1 + \tfrac{2}{5}P_3.$$A consequence: \( \int_{-1}^{1}x^kP_n\,dx = 0 \) whenever \( k < n \).
Associated Legendre functions \( P_l^m(x) = (1 - x^2)^{\lvert m\rvert/2}\dfrac{d^{\lvert m\rvert}}{dx^{\lvert m\rvert}}P_l(x) \) (up to a sign convention) solve the general angular equation and, with \( e^{im\phi} \), make the spherical harmonics \( Y_l^m(\theta, \phi) \). They exist only for \( \lvert m\rvert \le l \).
Hermite polynomials
First few (physicists’ convention): \( H_0 = 1 \), \( H_1 = 2x \), \( H_2 = 4x^2 - 2 \), \( H_3 = 8x^3 - 12x \), \( H_4 = 16x^4 - 48x^2 + 12 \).
Generating function and Rodrigues formula:
$$e^{2xt - t^2} = \sum_{n=0}^{\infty}H_n(x)\frac{t^n}{n!}, \qquad H_n(x) = (-1)^ne^{x^2}\frac{d^n}{dx^n}e^{-x^2}$$Orthogonality:
$$\boxed{\int_{-\infty}^{\infty}e^{-x^2}H_m(x)H_n(x)\,dx = 2^nn!\sqrt{\pi}\,\delta_{mn}}$$Recurrences: \( H_{n+1} = 2xH_n - 2nH_{n-1} \) and \( H_n' = 2nH_{n-1} \).
Parity: \( H_n(-x) = (-1)^nH_n(x) \). The oscillator wavefunctions are \( \psi_n \propto H_n(\xi)e^{-\xi^2/2} \) (see One-Dimensional Problems); the series for Hermite’s equation terminates only when the parameter is an integer \( n \), which is the quantisation \( E_n = (n + \tfrac{1}{2})\hbar\omega \).
Laguerre polynomials
First few: \( L_0 = 1 \), \( L_1 = 1 - x \), \( L_2 = \tfrac{1}{2}(x^2 - 4x + 2) \), \( L_3 = \tfrac{1}{6}(-x^3 + 9x^2 - 18x + 6) \). Every \( L_n(0) = 1 \).
Generating function and Rodrigues formula:
$$\frac{e^{-xt/(1-t)}}{1 - t} = \sum_{n=0}^{\infty}L_n(x)\,t^n, \qquad L_n(x) = \frac{e^x}{n!}\frac{d^n}{dx^n}\left(x^ne^{-x}\right)$$Orthogonality: \( \displaystyle\int_0^\infty e^{-x}L_m(x)L_n(x)\,dx = \delta_{mn} \).
Recurrence: \( (n + 1)L_{n+1} = (2n + 1 - x)L_n - nL_{n-1} \).
The associated Laguerre polynomials \( L_n^k(x) = (-1)^k\dfrac{d^k}{dx^k}L_{n+k}(x) \) appear in the hydrogen radial functions. Normalisations and index conventions for \( L_n^k \) differ between textbooks (some older books use a different definition), so check which one a question uses. What does not depend on convention: the hydrogen radial function \( R_{nl} \) has \( n - l - 1 \) nodes.
Bessel functions
Series:
$$J_\nu(x) = \sum_{k=0}^{\infty}\frac{(-1)^k}{k!\,\Gamma(k + \nu + 1)}\left(\frac{x}{2}\right)^{2k+\nu}$$so \( J_0(0) = 1 \), \( J_n(0) = 0 \) for \( n \ge 1 \), and for small \( x \), \( J_n(x) \approx (x/2)^n/n! \). The second solution \( Y_\nu \) (Neumann function) diverges at \( x = 0 \), so it is dropped in any problem that includes the axis.
Generating function (integer order):
$$e^{(x/2)(t - 1/t)} = \sum_{n=-\infty}^{\infty}J_n(x)\,t^n$$which gives \( J_{-n}(x) = (-1)^nJ_n(x) \) for integer \( n \).
Recurrences:
$$J_{\nu-1} + J_{\nu+1} = \frac{2\nu}{x}J_\nu, \qquad J_{\nu-1} - J_{\nu+1} = 2J_\nu'$$and their combinations
$$\frac{d}{dx}\left[x^\nu J_\nu\right] = x^\nu J_{\nu-1}, \qquad \frac{d}{dx}\left[x^{-\nu}J_\nu\right] = -x^{-\nu}J_{\nu+1}.$$With \( \nu = 0 \): \( J_0' = -J_1 \); with \( \nu = 1 \): \( (xJ_1)' = xJ_0 \).
Half-integer order: elementary functions.
$$J_{1/2}(x) = \sqrt{\frac{2}{\pi x}}\sin x, \qquad J_{-1/2}(x) = \sqrt{\frac{2}{\pi x}}\cos x$$These lead to the spherical Bessel functions, e.g. \( j_0(x) = \sin x/x \).
Zeros and large \( x \): \( J_\nu(x) \approx \sqrt{2/\pi x}\cos(x - \nu\pi/2 - \pi/4) \), so successive zeros approach a spacing of \( \pi \). The first zeros are \( J_0 \): 2.405, 5.520; \( J_1 \): 3.832. These set the normal modes of a circular membrane fixed at its rim: \( J_m(\alpha_{mn}r/a) \).
Orthogonality on \( [0, a] \): for zeros \( \alpha_{\nu m} \), \( \alpha_{\nu n} \) of \( J_\nu \),
$$\int_0^a x\,J_\nu\!\left(\frac{\alpha_{\nu m}x}{a}\right)J_\nu\!\left(\frac{\alpha_{\nu n}x}{a}\right)dx = \frac{a^2}{2}\left[J_{\nu+1}(\alpha_{\nu n})\right]^2\delta_{mn}.$$The gamma function
\( \Gamma(z) = \int_0^\infty t^{z-1}e^{-t}\,dt \), with \( \Gamma(z + 1) = z\Gamma(z) \), \( \Gamma(n + 1) = n! \) and \( \Gamma(\tfrac{1}{2}) = \sqrt{\pi} \). So \( \Gamma(\tfrac{3}{2}) = \tfrac{\sqrt\pi}{2} \), \( \Gamma(\tfrac{5}{2}) = \tfrac{3\sqrt\pi}{4} \approx 1.329 \). It enters through the Bessel series and through every Gaussian moment: \( \int_0^\infty x^{n}e^{-x^2}dx = \tfrac{1}{2}\Gamma\!\left(\tfrac{n+1}{2}\right) \).
graph TD
A[Separate Laplace or Schrodinger equation] --> B{Geometry}
B -->|spherical, angle theta| C[Legendre P_l and associated P_l^m]
B -->|cylindrical, radius| D[Bessel J_nu]
B -->|oscillator potential| E[Hermite H_n]
B -->|Coulomb potential, radius| F[Associated Laguerre]
C --> G[Orthogonal: weight 1 on -1 to 1]
D --> H[Orthogonal: weight x on 0 to a]
E --> I[Orthogonal: weight exp of -x^2]
F --> J[Orthogonal: weight exp of -x on 0 to infinity]Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| \( P_n(1) = 1 \) for all \( n \) | \( L_n(0) = 1 \) for all \( n \) | |
| \( J_0(0) = 1 \) | \( J_n(0) = 0 \) for \( n \ge 1 \); \( Y_\nu \) diverges at 0 | |
| \( J_{-n} = (-1)^nJ_n \) (integer \( n \)): not independent | \( J_{-\nu} \), \( J_\nu \) independent for non-integer \( \nu \) | |
| Hermite (physicists’): \( H_2 = 4x^2 - 2 \) | Probabilists’ \( He_2 = x^2 - 1 \) | |
| Ordinary point: plain power series | Regular singular point: Frobenius series with index \( r \) | |
| Repeated root \( m \) (constant coefficients): \( xe^{mx} \) | Repeated root \( m \) (Euler–Cauchy): \( x^m\ln x \) |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper.
Level 1: Foundation
- Write \( P_2(x) \).
- What is \( \Gamma(5/2) \)?
- Classify the point \( x = 0 \) for Bessel’s equation.
Answers
- \( \tfrac{1}{2}(3x^2 - 1) \).
- \( 3\sqrt\pi/4 \approx 1.329 \). \( \Gamma(\tfrac{5}{2}) = \tfrac{3}{2}\cdot\tfrac{1}{2}\cdot\Gamma(\tfrac{1}{2}) \).
- Regular singular point. In standard form \( P = 1/x \), \( Q = 1 - \nu^2/x^2 \); \( xP = 1 \) and \( x^2Q = x^2 - \nu^2 \) are analytic.
Level 2: Part B Style
Practice Q1. A particular solution of \( y'' + 4y = \sin2x \) is:
(a) \( \tfrac{1}{8}\sin2x \) (b) \( -\tfrac{x}{4}\cos2x \) (c) \( \tfrac{x}{4}\sin2x \) (d) \( -\tfrac{1}{4}\cos2x \)
Answer: (b)
The forcing frequency equals the natural frequency, so the guess is \( x(A\cos2x + B\sin2x) \), giving \( A = -\tfrac{1}{4} \), \( B = 0 \). Check: \( y = -\tfrac{x}{4}\cos2x \) gives \( y'' = \sin2x + x\cos2x \), and \( y'' + 4y = \sin2x \). Options (a) and (d) solve the homogeneous equation, so they give 0 on the right; (c) gives \( \cos2x \) on the right.
Practice Q2. The Wronskian of two independent solutions of \( x^2y'' - 2xy' + 2y = 0 \) (\( x > 0 \)) is proportional to:
(a) \( x^{-2} \) (b) \( 1 \) (c) \( x^2 \) (d) \( e^{x^2} \)
Answer: (c) x²
Standard form: \( y'' - \tfrac{2}{x}y' + \tfrac{2}{x^2}y = 0 \), so \( P = -2/x \) and \( W = C\exp(\int2/x\,dx) = Cx^2 \). Check: the equation is Euler–Cauchy with \( m(m-1) - 2m + 2 = 0 \), so \( m = 1, 2 \); \( W(x, x^2) = x\cdot2x - x^2\cdot1 = x^2 \). Option (d) comes from not dividing by \( x^2 \).
Practice Q3. The value of \( \int_{-1}^{1}\left[P_3(x)\right]^2dx \) is:
(a) \( 2/3 \) (b) \( 2/5 \) (c) \( 2/7 \) (d) 1
Answer: (c) 2/7
\( \int_{-1}^{1}P_n^2\,dx = 2/(2n + 1) \) with \( n = 3 \). Option (d) assumes the polynomials are normalised; they are normalised to \( P_n(1) = 1 \), not to unit norm.
Practice Q4. In the expansion \( x^3 = a_1P_1(x) + a_3P_3(x) \), the coefficients are:
(a) \( a_1 = \tfrac{3}{5},\ a_3 = \tfrac{2}{5} \) (b) \( a_1 = \tfrac{2}{5},\ a_3 = \tfrac{3}{5} \) (c) \( a_1 = 1,\ a_3 = 0 \) (d) \( a_1 = \tfrac{1}{3},\ a_3 = \tfrac{2}{3} \)
Answer: (a)
\( P_3 = \tfrac{1}{2}(5x^3 - 3x) \Rightarrow x^3 = \tfrac{2}{5}P_3 + \tfrac{3}{5}x \). Check at \( x = 1 \): \( \tfrac{3}{5} + \tfrac{2}{5} = 1 \), as required since every \( P_n(1) = 1 \). Option (d) is the expansion of \( x^2 \) in \( P_0, P_2 \).
Level 3: Part C Style
Practice Q5. Solve \( y'' + 2y' + 5y = 0 \) with \( y(0) = 1 \), \( y'(0) = 1 \). Which of the following are correct?
- The motion is underdamped with angular frequency 2.
- \( y(t) = e^{-t}(\cos2t + \sin2t) \).
- \( y(\pi/2) \approx -0.208 \).
- The successive zeros of \( y \) are separated by \( \pi \).
(a) 1 and 2 only (b) 1, 2 and 3 only (c) 2, 3 and 4 only (d) all four
Answer: (b) 1, 2 and 3 only
Statement 1: \( m^2 + 2m + 5 = 0 \Rightarrow m = -1 \pm 2i \). True.
Statement 2: \( y = e^{-t}(A\cos2t + B\sin2t) \), \( y(0) = A = 1 \), \( y'(0) = -A + 2B = 1 \Rightarrow B = 1 \). True.
Statement 3: \( y(\pi/2) = e^{-\pi/2}(\cos\pi + \sin\pi) = -e^{-\pi/2} \approx -0.208 \). True.
Statement 4: \( \cos2t + \sin2t = \sqrt2\cos(2t - \pi/4) \) vanishes every half-period of \( \cos2t \), i.e. every \( \pi/2 \). False; the separation is \( \pi/2 \).
Practice Q6. Which of the following statements about Legendre polynomials are correct?
- \( \int_{-1}^{1}P_2(x)\,dx = 0 \).
- \( \int_{-1}^{1}xP_1(x)P_2(x)\,dx = \tfrac{4}{15} \).
- \( P_3(0) = -\tfrac{1}{2} \).
- \( \sum_{n=0}^{\infty}P_n(1)\,t^n = \dfrac{1}{1 - t} \) for \( \lvert t\rvert < 1 \).
(a) 1 and 2 only (b) 1, 2 and 4 only (c) 2, 3 and 4 only (d) all four
Answer: (b) 1, 2 and 4 only
Statement 1: \( \int P_2 = \int P_2P_0 = 0 \) by orthogonality. True.
Statement 2: \( xP_1 = x^2 = \tfrac{1}{3}P_0 + \tfrac{2}{3}P_2 \), so the integral is \( \tfrac{2}{3}\cdot\tfrac{2}{5} = \tfrac{4}{15} \). True.
Statement 3: \( P_3 \) is odd, so \( P_3(0) = 0 \). (\( -\tfrac{1}{2} \) is \( P_2(0) \).) False.
Statement 4: at \( x = 1 \) the generating function is \( (1 - 2t + t^2)^{-1/2} = 1/(1 - t) \), consistent with \( P_n(1) = 1 \). True.
Practice Q7. Which of the following are correct for Bessel functions of the first kind?
- \( J_{1/2}(x) = \sqrt{2/(\pi x)}\,\sin x \).
- \( \dfrac{d}{dx}\left[xJ_1(x)\right] = xJ_0(x) \).
- \( J_0'(x) = -J_1(x) \).
- \( J_{-1}(x) = J_1(x) \).
(a) 1 and 3 only (b) 1, 2 and 3 only (c) 2, 3 and 4 only (d) all four
Answer: (b) 1, 2 and 3 only
Statement 1: the series with \( \nu = \tfrac{1}{2} \) sums to \( \sqrt{2/\pi x}\sin x \). True.
Statement 2: \( \tfrac{d}{dx}[x^\nu J_\nu] = x^\nu J_{\nu-1} \) with \( \nu = 1 \). True.
Statement 3: \( \tfrac{d}{dx}[x^{-\nu}J_\nu] = -x^{-\nu}J_{\nu+1} \) with \( \nu = 0 \). True.
Statement 4: \( J_{-n} = (-1)^nJ_n \), so \( J_{-1} = -J_1 \). False.
Quick Revision Table
| Item | Result |
|---|---|
| Integrating factor | \( \mu = e^{\int P} \); \( y = \mu^{-1}\left[\int\mu Q + C\right] \) |
| Abel | \( W = Ce^{-\int P} \) (standard form) |
| Damping | critical at \( b = 2\sqrt{mk} \); \( \omega_d = \sqrt{\omega_0^2 - \gamma^2} \) |
| Resonance | multiply the guess by \( x \) |
| Euler–Cauchy | \( y = x^m \), \( m(m-1) + am + b = 0 \) |
| Frobenius | indicial \( r(r-1) + p_0r + q_0 = 0 \) |
| Legendre | \( \int P_mP_n = \tfrac{2}{2n+1}\delta_{mn} \); \( P_n(1) = 1 \); \( (1 - 2xt + t^2)^{-1/2} \) |
| Hermite | \( \int e^{-x^2}H_mH_n = 2^nn!\sqrt\pi\,\delta_{mn} \); \( H_{n+1} = 2xH_n - 2nH_{n-1} \) |
| Laguerre | \( \int_0^\infty e^{-x}L_mL_n = \delta_{mn} \); \( L_n(0) = 1 \) |
| Bessel | \( (x^\nu J_\nu)' = x^\nu J_{\nu-1} \); \( J_0' = -J_1 \); zeros of \( J_0 \): 2.405, 5.520 |
| Gamma | \( \Gamma(\tfrac12) = \sqrt\pi \), \( \Gamma(n+1) = n! \) |
Key Takeaways
What’s Next
- Fourier Series, Fourier and Laplace Transforms: the Laplace transform turns these initial-value problems into algebra.
- Linear Algebra, Matrices and Eigenvalues, the previous topic.
- Sturm–Liouville Problems and Green’s Function in Mathematical Sciences, for why these families are orthogonal.
- Angular Momentum, Spin and the Hydrogen Atom, where Legendre and Laguerre functions become orbitals.