Solving an implicit function, fsolve vs. fzero
Show older comments
Hi,
I'm trying to solve an implicit function for x that looks like this:
A*x^q + Bx - C = 0
The solution is within (0,1). Speed is a huge issue. The solution is used to write down a new problem. (A and B both depend on the previous value of x). Both fsolve and fzero are too slow. Any ideas?
Thanks.
4 Comments
Geoff
on 4 Jul 2012
have you tried fminsearch?
Walter Roberson
on 5 Jul 2012
What are the knowns and unknowns ? Are you only solving for x, or for x, q, and C ? Is q a positive integer or can it be fractional ?
Richard Brown
on 5 Jul 2012
Is there known to be only one solution in the interval? Is the solution to a subsequent problem known to be close to the solution to the current one?
darius
on 5 Jul 2012
Answers (2)
Anton Semechko
on 5 Jul 2012
You can try the following approach:
abc=rand(3,1); % A, B, C parameters
q=8; % order of the polynomial
Nmax=1E3; % maximum number of iterations
f_tol=1E-6; % convergence tolerance
k=0.9; % relaxation parameter, must be in the range (0,1]
x=0.5; % starting point
f=abc(1)*x^q + abc(2)*x - abc(3); % initial function value
% search for the zero
for i=1:Nmax
x=k*x+(1-k)*(x-f); % update x
f=abc(1)*x^q + abc(2)*x - abc(3);
disp([i f x])
if abs(f)<f_tol, break; end
end
clear q Nmax f_tol k i
I tested the code given above for integer value of q in the range [1,10], and was able to find a solution in at most ~200 iterations. You can also play around with the k parameter to tune the rate of convergence. Note that setting k too low will cause the search to become unstable.
3 Comments
Richard Brown
on 5 Jul 2012
I think if you go with a derivative free approach a binary search would be quicker - it would get you 6 digits of accuracy guaranteed in 20 steps
Anton Semechko
on 5 Jul 2012
you can get similar performance if you decrease 'k' from 0.9 to a lower value (e.g. 0.6)
Richard Brown
on 5 Jul 2012
Actually, what I said doesn't work anyway! Too early in the morning, no coffee yet :)
Richard Brown
on 5 Jul 2012
Edited: Richard Brown
on 5 Jul 2012
How about doing a fixed number of Newton steps for each problem? (not necessarily doing a convergence check)
For example:
A, B, c, q = ...
nSteps = 5;
x = 0.5
for i = 1:nProblems
% Use x from previous problem to update
for i = 1:nSteps
x = x - (A*x^q + B*x + c) / (q*A*x^(q-1) + B);
end
% Update A, B, c, q to next problem
end
You could always add in a convergence check, or track the errors if you need to. Of course whether or not this works depends on the parameters of your problem, but with a bit of luck it will.
Categories
Find more on Optimization in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!