The Application Of Quick Group Search Optimizer To The Large-Scale Truss Structural Optimization

Truss structures are widely used structural types because of their good properties, such as rational stress mechanism, simple computation, convenient installation, strong applicability and light weight. Optimal design of large-scale truss structures is. a difficult problem with many design variables, where the search space has a large size and a great number of design constraints must be satisfied. Moreover, the relationships between design variables and design constraints as well as optimization objective are implicitly highly-nonlinear. The traditional gradient-based optimization algorithms are failed to solve this kind of problems effectively. Heuristic algorithms do not need special information of the problems such as computing of gradient. They are robust, general used, easy to implement and perform well in global optimization. At present the heuristic algorithms have been widely used in the field of structural optimization design.The Quick Group Search Optimizer (QGSO) is one of the recently presented heuristic algorithms. In this paper, the QGSO was tested by a 144-bar space truss, which was the scaled structure basd on the 72-bar space truss benchmark problem, to investigate the influence of the four major parameters in the QGSO on the optimal result. After that, a Modified Quick Group Search Optimizer (MQGSO) was presented and applied to ten large-scale truss structure optimization problems with continuous variables or discrete variables. The results indicate that the proposed MQGSO is suitable for solving this kind of optimization problems. It is capable of locating improved solutions compared with the other several heuristic algorithms used in the previous literatures.In consideration of the MQGSO’s lack of ability of exploitation compared to the QGSO in one of the numerical examples, the concept of cellular automata was introduced into the MQGSO and a Cellular Automata-Modified Quick Group Search Optimizer (CA-MQGSO) was presented, which with the improved ability of exploitation during local search. The CA-MQGSO was applied to the 693-bar braced barrel vault and achieved the expected effect. Further research will be carried out in the follow-up work.