A Simple Monte Carlo optimizer based on Adaptive Coordinate Sampling

This paper introduces a novel approach to optimize non-linear complex functions. The proposed algorithm is based on four key ideas: first, the optimization of one component of the current solution each time; second, the use of a truncated normal distribution as a random global optimization technique for optimizing the current dimension of the current solution; third, the evolution of the standard deviation of the sampling distribution in each iteration, as a mechanism of self-adaptation; and fourth, the restart of the algorithm for escaping of local optima. We test our approach using eight well-known benchmark problems. Our algorithm is comparable with, and, in some cases, better than, other well-established heuristic algorithms as evolution strategies and differential evolution, when considering the quality of the solutions obtained.