# Linear 1st Order PDE Query

• May 27th 2011, 12:08 PM
bugatti79
Linear 1st Order PDE Query
Folks,

Looking at past exam papers. There is one section as follows which I have no idea on.

Prove that a characteristic curve is parallel to a solution surface and hence explain why a characteristic curve that intersects any intial data must be a solution curve.

(Thinking)

Thanks
• Jun 3rd 2011, 09:57 PM
Jose27
Basically a characteristic is a curve along which the directional derivative is known and given by the coefficients. To get these we solve an ODE, and if we are given initial data (Cauchy problem) we have initial conditions to impose, so everything is determined. On the other hand if you want the general solution this initial conditions are missing so you get a family of curves all satisfying the same differential equation, and if this curve intersects a solution surface, by fixing this point and using the existence and uniqueness theorem for ODE, we get that there is an 'interval' around this point in the curve which is contained in the surface; we can then conclude that the whole characteristic is contained in the surface as long as both exist.
• Jun 4th 2011, 07:35 AM
bugatti79
Quote:

Originally Posted by Jose27
Basically a characteristic is a curve along which the directional derivative is known and given by the coefficients. To get these we solve an ODE, and if we are given initial data (Cauchy problem) we have initial conditions to impose, so everything is determined. On the other hand if you want the general solution this initial conditions are missing so you get a family of curves all satisfying the same differential equation, and if this curve intersects a solution surface, by fixing this point and using the existence and uniqueness theorem for ODE, we get that there is an 'interval' around this point in the curve which is contained in the surface; we can then conclude that the whole characteristic is contained in the surface as long as both exist.

Thanks for the description. I wonder was the question loooking for an algebraic method etc?

Thanks
• Jun 9th 2011, 09:23 PM
Jose27
Let's refrase the claim to something that may be more familiar (and hopefully that you have seen proved, or can prove): If we are given a Cauchy problem along a characteristic curve (ie. the transversality condition is violated) then this problem has either no solution or an infinite number of solutions.