Hey guys, I need a hint on how to effectivley solve a nonlinear optimization problem with NLopt package when my objective function contains some complex calculations and my nonlinear constraint contains the same complex calculations, something like the following: using NLopt function myfunc(x::Ve…

Which one is the way to go accroding to you opinion? The approach that I posted the code example for or the approach with global variables? I am asking because I am unsure whether global variables will really make it at all. I mean I am not familiar to the internals of NLopt, but I see a potential…

I am not an expert but I think the way to go is caching the result of the last calculation in two variables, x_cached and y_cached. Then before performing the expensive computation, you check whether the vector x is the same as x_cached for which you have already computed Y and stored it into y_cach…

That looks like a smart solution. Thanks for your support!

[image] mzaffalon: I was surprised to see that the objective function is computed first and then the constraints, but this may have to do with the choice of the algorithm. IIRC, NLopt always evaluates the objective before the constraints.

In my ignorance, I assumed that one may want to check whether the solution is feasible before computing the objective value.

Even if the point is infeasible, the objective value is still important to determine the next optimization step.

The risk is to sample a point that does not belong to the objective function domain, e.g. a negative number when f(x) = sqrt(x).

[image] mzaffalon: The risk is to sample a point that does not belong to the objective function domain, e.g. a negative number when f(x) = sqrt(x) . NLopt guarantees that the function will never be evaluated outside the feasible set for box constraints, but not for arbitrary nonlinear constra…

[image] stevengj: This is true for most nonlinear-programming algorithms that I’m aware of. I think you are right, but I am wondering if some simple heuristic could remedy this. Eg if the feasible set D is convex, x \in D, and y \notin D, one could find a z = \alpha x + (1 - \alpha) y \in D u…

NLopt - Same complex clalculations in objetive function and nonlinear constraints

General Usage Performance

mzaffalon June 25, 2020, 6:34am 2

Yes, it looks like it is the way to go.

Topic		Replies	Views
Optimising a non-linear user-defined function using NLopt Optimization (Mathematical) optimization	5	1102	May 14, 2021
NLopt not optimising General Usage optimization	13	1140	December 14, 2020
NLopt optimizer just return the initial value Optimization (Mathematical)	3	309	August 16, 2023
Nonlinear constrained optimization without gradient Optimization (Mathematical) question , optimization	3	662	January 23, 2023
How to use ForwardDiff with NLopt? Optimization (Mathematical) optimization , forwarddiff	9	2127	May 15, 2022

NLopt - Same complex clalculations in objetive function and nonlinear constraints

Related topics