
Re: fgoalattain: 3 GoalsExact to Achieve and 1 function to minimize
Posted:
Jan 2, 2013 10:37 AM


On 12/27/2012 3:20 PM, Allamarein wrote: >> Perhaps I should use 'nonlcon' and modify MyFun so that: >> >> >> >> Function [c,ceq] = Const(x,goal) >> >> [f4, f1_3] = MyFun(x) >> >> % f1_3 = y(1:3) >> >> % f4 = y(4) >> >> ceq = f1_3  goal >> >> end >> >> >> >> And minimize f4. However any evaluation of my function, in actual a double computational effort. I would avoid this 'brute force' approach >> >> What's your opinion? > The new question is: do you know an approach to avoid a double computational effort? >
There is relatively new documentation on how to avoid double computation: http://www.mathworks.com/help/optim/ug/writingconstraints.html#btfdwb7
Also, there is an example showing how to use exact goal attainment versus nonexact attainment: http://www.mathworks.com/help/optim/examples/multiobjectivegoalattainmentoptimization.html
Alan Weiss MATLAB mathematical toolbox documentation

