1

Little of the following may make sense to you, but I hope it does. I am starting out with PDEs, and there are particular 'tricks' in my notes that are used to reduce the dimensions of a PDE. This is done through variable substitution.

My question concerns the finding of new 'derivative' terms? Not sure what they are called. Here is an example.

Suppose I have the following term in my PDE: $\frac{\partial U}{\partial t}$ ($t$ stands for time), and I am making a following substitution: $\tau = T-t$, where $T$ is just a constant. Clearly: $\frac{\partial}{\partial t}=\frac{\partial \tau}{\partial t} \frac{\partial}{\partial \tau}$ which leads to the following: $\frac{\partial}{\partial t} = - \frac{\partial}{\partial \tau}$, because $\frac{\partial \tau}{\partial t} = -1$. But then I struggle to process the following.

Now suppose there is translation of the co-ordinate system such that we are transforming from $(\eta,\tau)$ to $(x,\tau^{'})$. Now, in my notes I have that:

$$\frac{\partial}{\partial \tau} = \frac{\partial x}{\partial \tau} \frac{\partial}{\partial x} + \frac{\partial \tau^{'}}{\partial \tau} \frac{\partial}{\partial \tau^{'}}$$

where $x=\eta+(r-\frac{1}{2} \sigma^2)\tau$ and $\tau=\tau^{'}$

And I do not follow why we suddenly have additional terms in that partial derivative expression. Is that because both changes of variable contain $\tau$?

Then why is this the case, there is no $\eta$ in both expressions:

$$\frac{\partial}{\partial \eta} = \frac{\partial x}{\partial \eta} \frac{\partial}{\partial x} + \frac{\partial \tau^{'}}{\partial \eta} \frac{\partial}{\partial \tau^{'}}$$

leonbloy
  • 66,702
naz
  • 3,516
  • 1
    Oh, I believe that $\frac{\partial \tau^{'}}{\partial \eta}$ term just vanishes – naz Dec 15 '15 at 17:33

1 Answers1

0

You don’t “suddenly have additional terms.” The expression for ${\partial\over\partial\tau}$ in your second example is simply an application of the chain rule for multivariable functions. The reason there is only one term in the expression for ${\partial\over\partial\eta}$ is, as i squared pointed out, that ${\partial\tau'\over\partial\eta}=0$, so the second term drops out. That’s the same reason that there’s only one term in the expression for ${\partial\over\partial t}$ in your first example—the other terms in the full expression vanish because it depends on only one of the variables.

amd
  • 55,457