期刊
ENGINEERING WITH COMPUTERS
卷 35, 期 4, 页码 1269-1296出版社
SPRINGER
DOI: 10.1007/s00366-018-0662-y
关键词
Natural frequency; Truss optimization; Meta-heuristics; Adaptive mechanism; Exploration; Exploitation
The structural dynamic response predominantly depends upon natural frequencies which fabricate these as a controlling parameter for dynamic response of the truss. However, truss optimization problems subjected to multiple fundamental frequency constraints with shape and size variables are more arduous due to its characteristics like non-convexity, non-linearity, and implicit with respect to design variables. In addition, mass minimization with frequency constraints are conflicting in nature which intricate optimization problem. Using meta-heuristic for such kind of problem requires harmony between exploration and exploitation to regulate the performance of the algorithm. This paper proposes a modification of a nature inspired Symbiotic Organisms Search (SOS) algorithm called a Modified SOS (MSOS) algorithm to enhance its efficacy of accuracy in search (exploitation) together with exploration by introducing an adaptive benefit factor and modified parasitism vector. These modifications improved search efficiency of the algorithm with a good balance between exploration and exploitation, which has been partially investigated so far. The feasibility and effectiveness of proposed algorithm is studied with six truss design problems. The results of benchmark planar/space trusses are compared with other meta-heuristics. Complementarily the feasibility and effectiveness of the proposed algorithms are investigated by three unimodal functions, thirteen multimodal functions, and six hybrid functions of the CEC2014 test suit. The experimental results show that MSOS is more reliable and efficient as compared to the basis SOS algorithm and other state-of-the-art algorithms. Moreover, the MSOS algorithm provides competitive results compared to the existing meta-heuristics in the literature.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据