[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Reverse function numerically

From: stn021
Subject: Re: Reverse function numerically
Date: Wed, 31 Jan 2018 16:49:28 +0100

Hello all,

thanks for the ideas.

I have been busy so it took me a while to get back to you.

So far I have checked out the hint of Stephen Montgomery-Smith, which
was fsolve().

That essentially does the same as I have previously done with
leasqr(), only a bit more efficient.

With leasqr() it takes about 150ms to calculate one single value. With
fsolve() it gets done in 90ms.
That is an improvement of course but I had hoped to find something
more significant.
For comparison: the "forward"-function takes about 5 microseconds per value.

@Juan Pablo Carbajal
The function in my example is not the actual function but it has a
very similar shape.
Specifically the two intersecting (almost-)planes are similar and the
limit of all variables in the range of 0..1 is also given.
So I believe that any solutions for the example would work on the
original functions as well
The actual modeling-function involves lots of sums of nonlinear
functions and would IMO not help at this point.

I continue work on the matter ...

THX for now, Stefan

reply via email to

[Prev in Thread] Current Thread [Next in Thread]