Re: Is it possible?

• To: mathgroup at smc.vnet.net
• Subject: [mg72965] Re: Is it possible?
• From: dh <dh at metrohm.ch>
• Date: Fri, 26 Jan 2007 07:46:43 -0500 (EST)
• References: <epa2ic\$iqn\$1@smc.vnet.net>

```Hi,
First, note that it does not matter how v depends on a internally. You
have a scalar function of one argument: y=f[x]. All you are asking for
is the derivative of the inverse function of a given function . The key
word here is "inverse function". The derivative of the inverse function
is the inverse of the derivative. Consider dy=f'[x] dx and you want
dx/dy=(dy/dx)^-1= 1/f'[x], This looks simple but there is a catch. The
formula gives you dx/dy as a function of x. If this is fine your are
done. However if you need a function of y you will need the inverse
function to f. In general this is hard, but you may be lucky and have a
simple case, all depends on the actual function. For the general case,
you can search for the key word "lagrage inversion theorem".
Further, obviously, inversion is impossible in a region where f' is zero
somewhere.

Daniel

Travelmate wrote:
> Hi,
>
> I'd really appreciate your help...
>
> I'been asked to perform a kind of analysis I've never done before..and
> I'm not even sure that it's possible.
>
> Suppose you have value function V=v[x(a,b,c,d),y(....),z(......)]
>
> The key-parameter is 'a' and it turns out that the first and second
> (partial) derivatives of V with respect to 'a' are negative and positive
> respectively. (V is continous)
>
> Now there's a problem: I've been asked to invert V(a) (keeping b,c and d
> constant)and to study the derivatives of this inverse function with
> respect to b, c and d. I'm not sure it's possible to do such a thing.
> ANd even if it is possible, I do not know how to start.
>
> Could you be so kind to tell me:
> 1) if it is actually possible to perform this study;
> 2) if the answer is yes, could you suggest me a starting point or any
> reference?
>