First-Order PDEs: Lagrange and Charpit Methods
CSIR NET PDE: linear, quasilinear and nonlinear first-order PDEs, complete, general and singular integrals, Lagrange's method, Charpit's method, Clairaut form.
- 24 min
On this page · 14 sections
Prerequisites
- First-order ODEs and exact differentials: Lagrange’s auxiliary equations are a system of ODEs, and Charpit ends with integrating \( dz = p\,dx + q\,dy \).
- Partial derivatives, the chain rule and Jacobians, as in Metric Spaces and Several Variables.
- Envelopes of one- and two-parameter families of surfaces (eliminate the parameters from the family and its parameter derivatives).
Notation used throughout the chapter: \( z = z(x, y) \), \( p = \frac{\partial z}{\partial x} \), \( q = \frac{\partial z}{\partial y} \).
The Hook: One Equation, Three Kinds of Answer
Classifying First-Order PDEs
The general first-order PDE in two independent variables is \( F(x, y, z, p, q) = 0 \). The classification depends on how \( p, q \) and \( z \) enter.
| Type | Form | Example |
|---|---|---|
| Linear | \( a(x, y)\,p + b(x, y)\,q + c(x, y)\,z = d(x, y) \) | \( xp + yq = 2z \) |
| Semilinear | \( a(x, y)\,p + b(x, y)\,q = c(x, y, z) \) | \( p + q = z^2 \) |
| Quasilinear | \( a(x, y, z)\,p + b(x, y, z)\,q = c(x, y, z) \) | \( zp + q = 0 \) |
| Nonlinear | anything else | \( p^2 + q^2 = 1 \), \( pq = z \) |
Linear \( \subset \) semilinear \( \subset \) quasilinear \( \subset \) general. “Quasilinear” means linear in the derivatives \( p, q \); the coefficients may depend on \( z \). Lagrange’s method handles the whole quasilinear class; Charpit’s method handles the general class.
Formation of PDEs
Eliminating arbitrary constants. A relation \( f(x, y, z, a, b) = 0 \) with two constants gives, with its \( x \)- and \( y \)-derivatives, three equations; eliminate \( a, b \) to get a first-order PDE. If there are more constants than independent variables, second-order derivatives are needed.
Eliminating an arbitrary function. From \( \varphi(u, v) = 0 \) with \( u, v \) known functions of \( x, y, z \), differentiate and eliminate \( \varphi \):
$$\frac{\partial(u, v)}{\partial(y, z)}\, p + \frac{\partial(u, v)}{\partial(z, x)}\, q = \frac{\partial(u, v)}{\partial(x, y)}.$$The result is always quasilinear. This is exactly why Lagrange’s method works backwards: its general solution is \( \varphi(u, v) = 0 \).
| Relation | Eliminate | PDE |
|---|---|---|
| \( z = ax + by + ab \) | \( a, b \) | \( z = px + qy + pq \) |
| \( z = (x - a)^2 + (y - b)^2 \) | \( a, b \) | \( p^2 + q^2 = 4z \) |
| \( z = f(x^2 + y^2) \) | \( f \) | \( yp - xq = 0 \) |
| \( z = x f\!\left(\frac{y}{x}\right) \) | \( f \) | \( xp + yq = z \) |
| \( z = f(x + ay) \) | \( f \) | \( q = ap \) |
Kinds of Solution
| Name | What it is | How to get it |
|---|---|---|
| Complete integral | a two-parameter family \( z = \varphi(x, y, a, b) \) solving the PDE, with \( a, b \) genuinely independent (rank of \( \begin{pmatrix} \varphi_a & \varphi_{xa} & \varphi_{ya} \\ \varphi_b & \varphi_{xb} & \varphi_{yb} \end{pmatrix} \) is 2) | Charpit, standard forms |
| General integral | the envelope of a one-parameter sub-family obtained by putting \( b = \psi(a) \) with \( \psi \) arbitrary | eliminate \( a \) from \( z = \varphi(x, y, a, \psi(a)) \) and \( \partial_a = 0 \) |
| Singular integral | the envelope of the whole two-parameter family | eliminate \( a, b \) from \( z = \varphi \), \( \varphi_a = 0 \), \( \varphi_b = 0 \) |
| Particular integral | one member: fixed \( a, b \) | substitute numbers |
For quasilinear equations the general integral is written directly as \( \varphi(u, v) = 0 \), and singular integrals are rare. For Clairaut equations the singular integral is the usual exam question.
Lagrange’s Method for Quasilinear Equations
$$\boxed{P(x, y, z)\,p + Q(x, y, z)\,q = R(x, y, z)}$$Geometry. The normal to an integral surface \( z = z(x, y) \) is \( (p, q, -1) \). The equation says \( (P, Q, R) \cdot (p, q, -1) = 0 \): the vector field \( (P, Q, R) \) is tangent to every integral surface. So integral surfaces are made of the integral curves of \( (P, Q, R) \), the characteristic curves.
Method.
- Write Lagrange’s auxiliary equations $$\frac{dx}{P} = \frac{dy}{Q} = \frac{dz}{R}.$$
- Find two independent integrals \( u(x, y, z) = c_1 \) and \( v(x, y, z) = c_2 \).
- The general solution is \( \varphi(u, v) = 0 \) with \( \varphi \) arbitrary, equivalently \( u = f(v) \) or \( v = g(u) \).
graph TD
A["Pp + Qq = R"] --> B["dx/P = dy/Q = dz/R"]
B --> C{"Can two ratios be
separated directly?"}
C -->|yes| D["Integrate the pair"]
C -->|no| E["Multipliers l, m, n with
lP + mQ + nR = 0"]
E --> F["l dx + m dy + n dz = 0
is exact: integrate"]
D --> G["u = c1, v = c2"]
F --> G
G --> H["General solution: phi(u, v) = 0"]Four techniques for the auxiliary equations
| Technique | When | Example |
|---|---|---|
| Separate a pair | one ratio involves only two variables | \( \frac{dx}{x} = \frac{dy}{y} \Rightarrow \frac{y}{x} = c_1 \) |
| Reuse the first integral | the second pair has a third variable you can eliminate with \( c_1 \) | substitute \( y = c_1 x \), integrate, then put \( c_1 = \frac{y}{x} \) back |
| Multipliers | find \( l, m, n \) with \( lP + mQ + nR = 0 \), \( l\,dx + m\,dy + n\,dz \) exact | \( (1, 1, 1) \) or \( (x, y, z) \) or \( (\frac{1}{x}, \frac{1}{y}, \frac{1}{z}) \) |
| Combine ratios | \( \frac{l\,dx + m\,dy + n\,dz}{lP + mQ + nR} \) equals each ratio | turns a hard pair into an exact pair |
Standard Lagrange results worth knowing by sight
| Equation | General solution |
|---|---|
| \( ap + bq = 0 \) (constants) | \( z = f(bx - ay) \) |
| \( ap + bq = c \) | \( z = \frac{c}{a}x + f(bx - ay) \) (\( a \ne 0 \)) |
| \( xp + yq = nz \) | \( z = x^n f\!\left(\frac{y}{x}\right) \) (Euler: homogeneous of degree \( n \)) |
| \( yp - xq = 0 \) | \( z = f(x^2 + y^2) \) |
| \( p + q = z \) | \( z = e^x f(x - y) \) |
| \( (y - z)p + (z - x)q = x - y \) | \( \varphi(x + y + z,\ x^2 + y^2 + z^2) = 0 \) |
Charpit’s Method for Nonlinear Equations
For \( F(x, y, z, p, q) = 0 \) not quasilinear, look for a second equation \( G(x, y, z, p, q, a) = 0 \) such that the two together give \( p \) and \( q \) as functions of \( x, y, z \) for which \( dz = p\,dx + q\,dy \) is integrable. That integrability is the compatibility of \( F \) and \( G \):
$$[F, G] = \frac{\partial(F, G)}{\partial(x, p)} + p\,\frac{\partial(F, G)}{\partial(z, p)} + \frac{\partial(F, G)}{\partial(y, q)} + q\,\frac{\partial(F, G)}{\partial(z, q)} = 0.$$This is a linear first-order PDE for \( G \); its Lagrange auxiliary equations are Charpit’s equations:
$$\boxed{\frac{dx}{F_p} = \frac{dy}{F_q} = \frac{dz}{p F_p + q F_q} = \frac{-dp}{F_x + p F_z} = \frac{-dq}{F_y + q F_z}}$$(Many books write the same thing as \( \frac{dp}{F_x + pF_z} = \frac{dq}{F_y + qF_z} = \frac{dz}{-pF_p - qF_q} = \frac{dx}{-F_p} = \frac{dy}{-F_q} \). The two are identical; only the signs of all denominators are flipped.)
Procedure.
- Compute \( F_x, F_y, F_z, F_p, F_q \) and write the five ratios.
- Find one simple integral involving \( p \) or \( q \) (or both), with constant \( a \). The \( dp \) and \( dq \) ratios are usually the place to look.
- Solve \( F = 0 \) and this integral for \( p \) and \( q \).
- Integrate \( dz = p\,dx + q\,dy \), adding a second constant \( b \). The result is a complete integral.
The four standard forms (Charpit made short)
These are special cases where the Charpit equations give the extra integral immediately.
| Form | Trick | Complete integral |
|---|---|---|
| I. \( f(p, q) = 0 \) | \( dp = 0 \): \( p = a \), \( q = b \) with \( f(a, b) = 0 \) | \( z = ax + by + c \), \( f(a, b) = 0 \) |
| II. \( f(z, p, q) = 0 \) | try \( z = \phi(u) \), \( u = x + ay \): \( p = \phi' \), \( q = a\phi' \) | solve the ODE \( f(\phi, \phi', a\phi') = 0 \) |
| III. \( f(x, p) = g(y, q) \) | set both sides \( = a \); solve \( p = P(x, a) \), \( q = Q(y, a) \) | \( z = \int P\,dx + \int Q\,dy + b \) |
| IV. Clairaut \( z = px + qy + f(p, q) \) | replace \( p \to a \), \( q \to b \) | \( z = ax + by + f(a, b) \) |
Clairaut’s equation and its singular integral
For \( z = px + qy + f(p, q) \) the complete integral is the family of planes \( z = ax + by + f(a, b) \). The singular integral is the envelope: eliminate \( a, b \) from
$$z = ax + by + f(a, b), \qquad x + f_a(a, b) = 0, \qquad y + f_b(a, b) = 0.$$Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| quasilinear: linear in \( p, q \), coefficients may contain \( z \) | linear: coefficients free of \( z \), \( z \) appears to power 1 | |
| general integral \( \varphi(u, v) = 0 \) (arbitrary function) | complete integral (two arbitrary constants) | |
| singular integral: envelope of the whole two-parameter family | general integral: envelope of a one-parameter sub-family | |
| Lagrange’s auxiliary system: three ratios | Charpit’s auxiliary system: five ratios | |
| \( \frac{dz}{pF_p + qF_q} \) (Charpit) | \( \frac{dz}{R} \) (Lagrange) | |
| singular integral is a solution, not in the complete integral | particular integral is a member of it |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper. In Level 3, as in Part C of the Mathematical Sciences paper, one or more options may be correct.
Level 1: Foundation
- Classify \( zp + q = 0 \): linear, semilinear, quasilinear or nonlinear?
- Write Lagrange’s auxiliary equations for \( x^2 p + y^2 q = z^2 \) and find its general solution.
- Write a complete integral of \( pq = 1 \).
Answers
- Quasilinear, not semilinear. It is linear in \( p, q \), but the coefficient of \( p \) contains \( z \).
- \( \frac{dx}{x^2} = \frac{dy}{y^2} = \frac{dz}{z^2} \). Integrating pairs: \( \frac{1}{x} - \frac{1}{y} = c_1 \), \( \frac{1}{y} - \frac{1}{z} = c_2 \). General solution \( \varphi\!\left(\frac{1}{x} - \frac{1}{y},\ \frac{1}{y} - \frac{1}{z}\right) = 0 \).
- Form I: \( p = a \), \( q = \frac{1}{a} \), so \( z = ax + \frac{y}{a} + b \).
Level 2: Part B Style
Practice Q1. The general solution of \( xp + yq = 2z \) is (\( f \) arbitrary):
(a) \( z = x^2 + f(y) \) (b) \( z = x^2 f\!\left(\frac{y}{x}\right) \) (c) \( z = f(xy) \) (d) \( z = 2x f\!\left(\frac{y}{x}\right) \)
Answer: (b)
\( \frac{dx}{x} = \frac{dy}{y} = \frac{dz}{2z} \) gives \( \frac{y}{x} = c_1 \) and \( \frac{z}{x^2} = c_2 \). So \( z = x^2 f\!\left(\frac{y}{x}\right) \): \( z \) is homogeneous of degree 2. Option (d) has degree 1.
Practice Q2. The singular integral of \( z = px + qy + p^2 + q^2 \) is:
(a) \( 4z = x^2 + y^2 \) (b) \( 4z + x^2 + y^2 = 0 \) (c) \( z = -xy \) (d) it has none
Answer: (b)
Complete integral \( z = ax + by + a^2 + b^2 \). Then \( x + 2a = 0 \), \( y + 2b = 0 \): \( a = -\frac{x}{2} \), \( b = -\frac{y}{2} \). So \( z = -\frac{x^2}{2} - \frac{y^2}{2} + \frac{x^2}{4} + \frac{y^2}{4} = -\frac{x^2 + y^2}{4} \).
Check: \( p = -\frac{x}{2} \), \( q = -\frac{y}{2} \); \( px + qy + p^2 + q^2 = -\frac{x^2 + y^2}{2} + \frac{x^2 + y^2}{4} = -\frac{x^2 + y^2}{4} = z \).
Practice Q3. For \( F(x, y, z, p, q) = p^2 x + q^2 y - z = 0 \), the Charpit ratio \( \frac{-dp}{F_x + pF_z} \) equals:
(a) \( \frac{-dp}{p^2 - p} \) (b) \( \frac{-dp}{p^2 + p} \) (c) \( \frac{-dp}{2px} \) (d) \( \frac{-dp}{p^2} \)
Answer: (a)
\( F_x = p^2 \) and \( F_z = -1 \), so \( F_x + pF_z = p^2 - p \). Option (c) is \( F_p \), which belongs under \( dx \).
Practice Q4. The general solution of \( p + q = z \) is (\( f \) arbitrary):
(a) \( z = e^{x + y} f(x - y) \) (b) \( z = e^{x} f(x - y) \) (c) \( z = f(x + y) \) (d) \( z = x + f(x - y) \)
Answer: (b)
\( dx = dy = \frac{dz}{z} \): \( x - y = c_1 \) and \( z e^{-x} = c_2 \). Option (a) fails: with \( z = e^{x+y}f(x - y) \), \( p + q = 2z \). Note that \( e^{y} f(x - y) \) is also correct (it equals \( e^{x} g(x - y) \) with \( g(t) = e^{-t} f(t) \)), but it is not offered.
Level 3: Part C Style
Practice Q5. Consider \( (y - z)\,p + (z - x)\,q = x - y \). Which statements are true?
(a) The general solution is \( \varphi(x + y + z,\ x^2 + y^2 + z^2) = 0 \). (b) Every integral surface of the form \( \varphi(x + y + z, x^2 + y^2 + z^2) = 0 \) is a surface of revolution about the line \( x = y = z \). (c) \( z = x \) is a solution on the whole plane. (d) \( z = -x - y \) is a solution.
Answer: (a), (b) and (d)
- (a) True. Multipliers \( (1, 1, 1) \): \( P + Q + R = 0 \), so \( x + y + z = c_1 \). Multipliers \( (x, y, z) \): \( x(y - z) + y(z - x) + z(x - y) = 0 \), so \( x^2 + y^2 + z^2 = c_2 \).
- (b) True. Each level set fixes the distance from the origin and the projection on \( (1, 1, 1) \): it is a union of circles centred on the line \( x = y = z \), in planes perpendicular to it.
- (c) False. \( p = 1 \), \( q = 0 \): the left side is \( y - x \), the right side \( x - y \); equal only on \( x = y \).
- (d) True. \( p = q = -1 \): left side \( -(y - z) - (z - x) = x - y \). It is the case \( x + y + z = 0 \) of (a).
Practice Q6. Consider \( F = px + qy - pq = 0 \). Which statements are true?
(a) Charpit’s equations contain \( \frac{dp}{p} = \frac{dq}{q} \). (b) \( 2az = (ax + y)^2 + b \) is a complete integral. (c) The equation is quasilinear. (d) \( z = x^2 + y^2 \) is a solution.
Answer: (a) and (b)
- (a) True. \( F_x + pF_z = p \), \( F_y + qF_z = q \), so \( \frac{-dp}{p} = \frac{-dq}{q} \).
- (b) True. It is \( z = \frac{(ax + y)^2}{2a} + b' \) with \( b' = \frac{b}{2a} \), the complete integral found above.
- (c) False. The product \( pq \) makes it nonlinear in the derivatives.
- (d) False. \( p = 2x \), \( q = 2y \): \( px + qy = 2(x^2 + y^2) \) but \( pq = 4xy \).
Practice Q7. Consider the Clairaut equation \( z = px + qy + \sqrt{1 + p^2 + q^2} \). Which statements are true?
(a) \( z = ax + by + \sqrt{1 + a^2 + b^2} \) is a complete integral. (b) The upper hemisphere \( x^2 + y^2 + z^2 = 1 \), \( z > 0 \), is its singular integral. (c) The singular integral is obtained from the complete integral by a suitable choice of \( a \) and \( b \). (d) The plane \( z = 1 \) is a solution.
Answer: (a), (b) and (d)
- (a) True. Clairaut form: replace \( p, q \) by constants.
- (b) True. Worked above: \( a = -xs \), \( b = -ys \), \( s = (1 - x^2 - y^2)^{-1/2} \) gives \( z = \sqrt{1 - x^2 - y^2} \).
- (c) False. Every member of the complete integral is a plane; the hemisphere is their envelope, not a member.
- (d) True. It is the member \( a = b = 0 \). It touches the hemisphere at \( (0, 0, 1) \).
Quick Revision Table
| Idea | Fact |
|---|---|
| Quasilinear | \( P(x,y,z)p + Q(x,y,z)q = R(x,y,z) \) |
| Lagrange | \( \frac{dx}{P} = \frac{dy}{Q} = \frac{dz}{R} \); two independent integrals \( u, v \); \( \varphi(u, v) = 0 \) |
| Multipliers | \( lP + mQ + nR = 0 \) and \( l\,dx + m\,dy + n\,dz \) exact |
| Charpit | \( \frac{dx}{F_p} = \frac{dy}{F_q} = \frac{dz}{pF_p + qF_q} = \frac{-dp}{F_x + pF_z} = \frac{-dq}{F_y + qF_z} \) |
| Compatibility | \( [F, G] = 0 \) |
| \( f(p, q) = 0 \) | \( z = ax + by + c \), \( f(a, b) = 0 \) |
| \( f(z, p, q) = 0 \) | \( z = \phi(x + ay) \) |
| \( f(x, p) = g(y, q) \) | both \( = a \) |
| Clairaut | \( z = ax + by + f(a, b) \); singular: envelope |
| Euler | \( xp + yq = nz \Rightarrow z = x^n f(y/x) \) |
Key Takeaways
What’s Next
- The Cauchy Problem for First-Order PDEs: choosing the one integral surface through a given curve, and when that is impossible.
- Classification of Second-Order PDEs: characteristics again, now for second-order equations.
- Sequences and Series of Functions for the uniform convergence used later with Fourier series.
- Partial Differential Equations: Quick Revision.