An efficient and reliable global optimization algorithm is proposed by combining the stochastic approach of a genetic algorithm (GA) and the deterministic approach of the filled function method. In the combined algorithm, the GA serves as a supplier of desirable starting points for the filled function method. The filled function method finds the point that is lower than the minimum previously found. By exploiting the features of both constituents the global optimum can be found more efficiently and more reliably. The combined algorithm is treated numerically for various functions available in the literature and desirable features are ascertained.