基于自适应遗传算法的钢筋混凝土桥墩抗震设计方法

为满足钢筋混凝土桥墩抗震设计中对构件承载力和延性的要求，常需依靠设计者经验，对配筋方案进行反复调整，这就在一定程度上增加了结构设计难度，降低了工作效率。为解决这一实际问题，对传统自适应遗传算法中交叉率和变异率的计算方法进行改进，通过当前适应度值动态调整交叉率和变异率，在此基础上提出基于自适应遗传算法的钢筋混凝土桥墩结构抗震设计方法，实现对符合条件设计解的自动随机定向搜索，提高求解的收敛速度，并增强算法的鲁棒性。在对每一配筋方案的抗弯承载力和位移延性系数的计算中，又提出基于自适应遗传算法的钢筋混凝土桥墩弯矩 曲率关系的计算方法，提高求解效率。通过设计算例，可见该方法快速有效，计算过程无需人工干预，实现钢筋混凝土桥墩结构抗震设计的人工智能化。

Seismic design method for reinforcement concrete#br# piers based on adaptive genetic algorithm

To meet the strength and ductility demands of members in the seismic design of RC piers, reinforcement schemes are often modified by means of the experiences of designers, which will increase the difficulty of structures design and reduce the work efficiency. To solve this problem, the calculation method of the crossover and mutation rate in the general adaptive genetic algorithm was improved. The crossover and mutation rate was adjusted dynamically as the values of the fitness and the seismic design method for reinforcement concrete piers based on the adaptive genetic algorithm was proposed. Auto random orientation search was realized to get the solutions that match the design conditions. The method will be helpful to generate the new population modes. It can also prevent the calculation from local convergence in searching solutions and improve the robustness of the algorithm. In calculation on the flexural capacity and displacement ductility coefficient of each population individual, the method of moment-curvature relationship of RC piers based on the adaptive genetic algorithm was proposed to improve the solution efficiency. From the design examples, it can be seen that the method in the paper is fast and effective. The manual intervention is not needed in the calculation and artificial intelligence on the seismic design of RC piers can be realized.