frontier_multiobj

find a Pareto frontier for multiple objective mixed integer linear optimisation with two objectives

Consider a multiobjective optimisation to minimise both f1(x) and f2(x). A point x is Pareto efficient if there are no other points x' which satisfy f1(x')<=f1(x) and f2(x')<=f2(x) without both inequalities equal. This function finds a set of Pareto efficient points for the linear multiobjective optimisation subject to specified integer constraints, linear inequalities and linear equality constraints.

NOTE: REQUIRES OPTIMISATION TOOLBOX.

A small example:

[x,v]=frontier_multiobj( [6,4,2,1; -5,-4,-3,-2], 1:4, [], [], [], [], zeros(4,1), ones(4,1));

display(x);

display(v); % expect [0 1 2 3 5 6 7 9 11 12 13;0 -2 -3 -5 -6 -7 -9 -10 -11 -12 -14]

plot(v(1,:),-v(2,:),'x')

For other examples, run frontier_multiobj() without any input arguments.