Nowadays a wide range of complex optimization problems are solved in a distributed way by edge devices. This paper presents a new method of optimization, based on metaheuristics independent runs, objective function calls blocks forming and assignment and efficient metaheuristic selection for the blocks forming and assignment. Compared with fixed volumes of objective function calls blocks processing, the positive effect on time costs and optimization problem results has been achieved. The method proposed decreases the time cost up to 10 times with the similar problem solving results, and increases the resulting values of criteria significantly with the time cost increase up to 4 times in comparison with the minimum objective functions calls block size.