A new original algorithm, CONDOR ("COnstrained, Non-linear, Direct, parallel Optimisation using trust Region method for high-computing load function"), was developed. The goal of this algorithm is to find the minimum of an objective function in the least number of function evaluations.
It is assumed that the dominant computing cost of the optimisation process is the time needed to evaluate the objective function (one evaluation can range from 2 minutes to 2 days). The algorithm will try to minimize the number of evaluations, at the cost of a huge amount of routine work. The derivatives of the objective function need not to be known. The only information needed about the objective function is a simple method (written in Fortran, C++) or a program (a Unix, Windows, Solaris executable) which can evaluate the objective function at a given point. The algorithm has been specially developed to be very robust against noise inside the evaluation objective function. This situation is very general, the algorithm can thus be applied on a vast number of situations.
CONDOR outperforms many commercial, high-end optimiser and is maybe the fastest optimiser in its category (fastest in terms of number of function evaluations). When several CPU's are used, the performances of CONDOR are unmatched. The experimental results open wide possibilities in the field of noisy and high-computing-load objective functions optimisation like, for instance, industrial shape optimisation based on CFD (computation fluid dynamic) codes (The Method project uses this kind of objective function) or PDE (partial differential equations) solvers.