中文摘要：

介绍基因表达式编程（GEP）算法的基本原理和在参数优化中的实现过程，并将该算法应用于桁架结构的优化设计。针对标准GEP容易陷入局部最优解，且收敛速度慢的缺陷，对算法引入回溯机制，用停滞前一代的精英个体替换当前种群中所有适应度最差的个体，使较优个体有更多机会向不同方向进化，扩大最优解的搜索空间。25杆空间桁架的截面优化设计结果证明：算法改进后，搜索效率得到明显提高，并通过72杆空间桁架算例，证明了该方法在结构优化中的可行性和有效性。

英文摘要：

The paper described the basic principle of Gene Expression Programming （GEP） algorithm and its realization process in parameter optimization, and then applied GEP to the truss structural optimization design. For the standard algorithm tends to be trapped in local optima and has the shortcoming of slow convergence, the backtracking mechanism was introduced, that is, to use the elite individual of the generation before stagnation to replace all worst individuals of the stagnation generation, in such a way to make the better individuals have more opportunities to evolve in different directions, thus the search space of optimal solution is expanded. The results of the 25-bar space truss sizing optimization show that the search efficiency is improved greatly, and the 72-bar space truss example proves that the method in the structural optimization is feasible and effective.