Multi-variable minimization

Hi,

I am trying to find the minimum variables which fits the function f(x)=0. I tried to use fmincon and fminsearch to find the minimum values of the variables. But the values I am getting are not consistent.

Is there any other way to get the estimation of the values which fit the equation.

f(x)=gx-((gm-x(1))/(gm*x(2)-x(3)) where gx and gm are known values and i need to estimate the min(x) which gives the function f(x)=0.

[Jarrod Rivituso]

Could you share with us the code you are using with fmincon or fminsearch?

Answer 1

[Matt Tearle]

What do you mean by "the minimum variables"? x is a vector, so what are you trying to minimize? All the x values? The norm of x, under some metric?

This works for me:

g = @(x) norm(x,Inf);
fmincon(g,rand(3,1),[],[],[],[],[],[],@fcnstrnt,optimset('Algorithm','interior-point'))

And fcnstrnt.m:

function [foo,ceq] = fcnstrnt(x)
gx = pi;
gm = 4.2;
ceq = gx-(gm-x(1))/(gm*x(2)-x(3));
foo = -1; % No inequality constraints on x