Parameter estimation by simulated annealing.

I have a data of function of multiple input variable and single output variable, like y=b*x1+ c*x2+d*x3+e*x4+f*x5 , data of input and output is available. how can I find a,b,c,d,e,f by simulated annealing.

I need to find all 'A' paremeters, using simulated annealing matlab code. somebody help.

Answer 1

[Walter Roberson]

Example:

x = rand(15, 6);
Y = randn(15,1);

fun = @(A) sum( (A(1)*x(:,1) + A(2)*x(:,2) + (A(3)*x(:,3).^2 + A(4))./x(:,4) + A(5)./x(:,5).^2 + A(6)*x(6) - Y).^2 );

A = simulannealbnd(fun, randn(1,6), -inf*ones(1,6), inf*ones(1,6));

I assumed here that you have multiple x/Y combinations, as the solution is obviously not unique if you only have a single Y.