Ant Colony Optimization (ACO) has been used as one of the popular meta-heuristic
algorithms in structural optimization. In this algorithm, the selected cross sections are chosen
according to a parameter called “probability ratio”. This parameter and the way to choose the cross
sections from a list of cross sections, are the most important points in the optimization process.
Though the Ant Colony algorithm has a special ability in achieving the optimal point, in some
cases in order to avoid local optima, the utilization of special techniques is needed. In the present
paper, the first aim is to use Harmony Search (HS) algorithm to increase the local search ability of
the ACO. In this way a combined algorithm, denoted by HACOHS, is obtained with special
abilities to achieve a global optimum. For this purpose, optimal design of skeletal structures such
as trusses and steel frames is considered using the HACOHS. However, in the process of
optimization by HACOHS method, several GA selections are employed at the cross section
selection stage. Utilizing the Tournament (HACOHS-T), Roulette wheel (HACOHS-Ro), and
Rank (HACOHS-Ra) methods it is found that the HACOHS-T is the most efficient of these
algorithms for optimal design of skeletal structures.