Construction of foundations or embankments on a soft/loose soil deposit causes major problems in terms of geotechnical engineering. The vertical loads transferred from the foundation can cause failure and/or extreme settlements in the soft/loose soil. Additional geotechnical precautions must be taken in order to prevent such problems. Construction of piles under foundations is a widely used method. Piles are defined as point bearing piles or friction piles depending on the embedded length of the pile in the stiff layer or rock. Generally it is preferred to construct pile groups under the foundations to reduce the effects of differential settlements and eccentricity. The interaction of a pile with the others in a group is defined as "group efficiency" and this interaction causes reduction in the load-bearing capacities of the piles. It is necessary to calculate the bearing capacity of a single pile correctly and estimate the optimum number of piles in the group to make a safe and economical design. In this paper, it is aimed to investigate the robustness of Grey Wolf optimization algorithm for optimization problems of pile groups under vertical loads. In order to compare the validity of Grey Wolf Optimization algorithm, Particle Swarm Optimization algorithm and Improved Harmony Search algorithm are used. The proposed methods are intended to help engineers to make fast, safe and economical designs for pile groups. In this study, only the bearing capacities and optimization of bored pile groups are discussed.