*This preview shows
pages
1–2. Sign up to
view the full content.*

This ** preview**
has intentionally

**sections.**

*blurred***to view the full version.**

*Sign up*
**Unformatted text preview: **for the planar heat equation is given by the linear superposition formula u ( t, x, y ) = 1 4 t integraldisplayintegraldisplay f ( , ) e [( x ) 2 +( y ) 2 ] / (4 t ) d d. (12 . 125) We can interpret the solution formula (12.125) as a two-dimensional convolution u ( t, x, y ) = F ( t, x, y ) f ( x, y ) (12 . 126) of the initial data with a one-parameter family of progressively wider and shorter Gaussian filters F ( t, x, y ) = F ( t, x, y ; 0 , 0) = 1 4 t e [ x 2 + y 2 ] / (4 t ) . (12 . 127) As in (8.51), such a convolution can be interpreted as a weighted averaging of the function, and has the effect of smoothing out the initial signal f ( x, y ). Example 12.14. If our initial temperature distribution is constant on a circular region, say u (0 , x, y ) = braceleftbigg 1 x 2 + y 2 < 1 , , otherwise , then the solution can be evaluated using (12.125), as follows: u ( t, x, y ) = 1 4 t integraldisplay integraldisplay D e [( x ) 2 +( y ) 2 ] / (4 t ) d d, where the integral is over the unit disk D = { 2 + 2 1 } . Unfortunately, the integral cannot be expressed in terms of elementary functions. On the other hand, numerical evaluation of the integral is straightforward. A plot of the resulting radially symmetric solution appears in Figure 12.8. One could also interpret this solution as the diffusion of an animal population in a uniform, isotropic environment, or bacteria in a similarly uniform large petri dish, that is initially confined to a circular region....

View Full
Document