Solving an implicit function, fsolve vs. fzero

8 ビュー (過去 30 日間)
darius
darius 2012 年 7 月 4 日
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 件のコメント
Richard Brown
Richard Brown 2012 年 7 月 5 日
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
darius 2012 年 7 月 5 日
Thank you for the comments. fminsearch is also too slow.
x is the only unknown. q is a fraction for sure. The solution to the subsequent problem is known to be greater than the solution tot he current one.

サインインしてコメントする。

回答 (2 件)

Anton Semechko
Anton Semechko 2012 年 7 月 5 日
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 件のコメント
Anton Semechko
Anton Semechko 2012 年 7 月 5 日
you can get similar performance if you decrease 'k' from 0.9 to a lower value (e.g. 0.6)
Richard Brown
Richard Brown 2012 年 7 月 5 日
Actually, what I said doesn't work anyway! Too early in the morning, no coffee yet :)

サインインしてコメントする。


Richard Brown
Richard Brown 2012 年 7 月 5 日
編集済み: Richard Brown 2012 年 7 月 5 日
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.

カテゴリ

Help Center および File ExchangeOptimization についてさらに検索

タグ

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by