Skip to content
Sahithyan's S3
1
Sahithyan's S3 — Differential Equations

Partial Differential Equations

A PDE involves partial derivatives of a dependent variable with respect to 2 or more independent variables.

A general PDE in implicit form is f(x,y,u,ux,uy,uxx,uxy,)=0f(x,y,u,u_x,u_y,u_{xx},u_{xy},\ldots) = 0.

Here:

  • x,yx,y - independent variables
  • u(x,y)u(x,y) - dependent variable
  • uxu_x - first partial derivative w.r.t. xx
  • uyu_y - first partial derivative w.r.t. yy
  • uxxu_{xx} - 2nd partial derivative w.r.t. xx
  • uyyu_{yy} - 2nd partial derivative w.r.t. yy
  • uxyu_{xy} - partial derivative of uxu_x w.r.t. yy
  • uyxu_{yx} - partial derivative of uyu_y w.r.t. xx

Highest derivative appearing in a PDE.

Dependent variable and all derivatives appear linearly. Won’t have: ux2u_x^2, sin(u)\sin(u), uuxxu\cdot u_{xx}.

When a PDE is not linear.

When highest-order derivative terms appear linearly. but coefficients may depend on lower-order derivatives or variables. A subset of non-linear PDEs.

Every term contains uu or its derivatives. There are no constants.

Not discussed for non-linear PDEs.

When a PDE is not homogeneous.

The PDE is written without solving for any specific derivative. Everything is inside one general function.

f(x,y,u,ux,uy,uxx,uxy,)=0f(x,y,u,u_x,u_y,u_{xx},u_{xy},\ldots) = 0

The PDE is solved for the highest-order derivatives.

The PDE is solved for only one highest-order derivative. Basically explicit form for the specific highest derivative you care about.

Used to solve PDEs by characteristic curves.

Second-order equations are central in heat flow, vibrations and potential theory. Their classification determines what solving technique is appropriate.

A second-order linear PDE with 1 dependent variable and 2 independent variables.

Auxx+Buxy+Cuyy+Dux+Euy+Fu+G=0A u_{xx} + B u_{xy} + C u_{yy} + Du_x + E u_y + F u + G = 0

Here AA to GG are functions of xx or yy or both or constants.

Classification depends on the principal part L(u)=Auxx+Buxy+CuyyL(u) = Au_{xx} + Bu_{xy} + Cu_{yy}. When A,B,CA, B, C are functions of xx and yy, the classification may differ across different points.

For a PDE in canonical form, its discriminant is:

D=B24ACD = B^2 - 4AC

When D<0D \lt 0.

Examples:

  • uxx+uyy=0u_{xx} + u_{yy} = 0 (Laplace Equation)

When D>0D \gt 0.

Examples:

  • uxxuyy=1u_{xx} - u_{yy} = 1

When D=0D = 0.

Examples:

  • ut=kuxxu_t = k u_{xx} (Heat Equation)