Truss optimization with natural frequency bounds using improved symbiotic organisms search

Many engineering structures are subjected to dynamic excitation, which may lead to undesirable vibrations. The multiple natural frequency bounds in truss optimization problems can improve dynamic behaviour of structures. However, shape and size variables with frequency bounds are challenging due to its characteristic, which is non-linear, non-convex, and implicit with respect to the design variables. As the main contribution, this work proposes an improved version of a recently proposed Symbiotic Organisms Search (SOS) called an Improved SOS (ISOS) to tackle the above-mentioned challenges. The main motivation is to improve the exploitative behaviour of SOS since this algorithm significantly promotes exploration which is a good mechanism to avoid local solution, yet it negatively impacts the accuracy of solutions (exploitation) as a consequence. The feasibility and effectiveness of ISOS is studied with six benchmark planar/space trusses and thirty functions extracted from the CEC2014 test suite, and the results are compared with other meta-heuristics. The experimental results show that ISOS is more reliable and efficient as compared to the basis SOS algorithm and other state-of-the-art algorithms.