How to constrain fmincon unknowns to itself

I am attempting to use fmincon to solve a problem of multiple unknowns.

unknowns = fmincon(@(x)fun(x),initial,[],[],[],[],lb,ub)

I need four of the unknowns to be constrained such that

(unknowns(1) + unknowns(2)) < (unknowns(3) + unknowns(4))

I cannot hardcode this constraint into the upper and lower bounds of the problem, I was wondering if there was a way to implement it with the other options / nonlinear constraints?

Thanks for any input.

Answer 1

[Alan Weiss]

Or maybe with linear constraints:

A = [1 1 -1 -1];
b = 0;
unknowns = fmincon(@(x)fun(x),initial,A,b,[],[],lb,ub)

What do the A and b arguments represent? Let x represent your unknowns argument.

A*x <= b

x(1) + x(2) -x(3) -x(4) <= 0

x(1) + x(2) <= x(3) + x(4)

unknowns = fmincon(@(x)fun(x),initial,[],[],[],[],lb,ub)

Alan Weiss

MATLAB mathematical toolbox documentation

[somebody]

Thank you, that helps a lot.