# Which algorithm does Mathematica use for “RandomSearch”?

Recently I have used Mathematica for minimization of certain functions and I have realized that even though it lasts much longer, the method "RandomSearch" functions much better than the other methods for minimization. I am going to use these results for a research article and I thought it would be better if I also indicate the method explicitly. Is it possible to know that?

According to section "Random Search" in this guide, NMinimize with method RandomSearch scatters points randomly over the search domain and applies the local search method FindMinimum with these methods as starting values. For unconstrained optimization problems, the default local method is BFGS (or L-BFGS?) (equivalent to calling FindMinimum with Method -> "QuasiNewton") and for constrained optimization problems, an interior point method is used per default (equivalent to calling FindMinimum with Method -> "InteriorPoint").
• Indeed, it is mentioned here that FindMinimum[] uses (L-)BFGS and interior point methods. – J. M. is away Oct 19 '18 at 11:44