论文部分内容阅读
近年来,由于智能化方法——人工神经网络和遗传算法所具有的种种优点,其理论和应用研究得到工程界较广泛的关注。针对目前土钉支护优化设计中计算工作量大和求解时间长的问题,提出了将神经网络与遗传算法结合进行求解的思想,利用神经网络学习算法建立起输入参数(优化设计变量)和输出参数(安全系数最小值)之间的非线性映射关系,当神经网络学习达到收敛条件时,从映射关系就极易获得遗传算法求解优化问题所需的对应于给定设计变量的安全系数最小值的近似值,以代替每次必须进行的最小安全系数求解。算例结果表明,采用神经网络与遗传算法结合进行土钉最小长度优化求解所需要的时间大大减少,而且具有良好的效果。
In recent years, due to the advantages of intelligent methods - artificial neural networks and genetic algorithms, their theoretical and applied research have attracted more and more attention in engineering field. Aiming at the problem of large computational workload and long solution time in the optimal design of soil nailing support, the idea of combining neural network and genetic algorithm is put forward, and the input parameters (optimal design variables) and output parameters (Minimum safety factor). When neural network learning reaches the convergence condition, the minimum value of the safety factor corresponding to a given design variable required by the genetic algorithm to solve the optimization problem can be easily obtained from the mapping relationship Approximation to replace the minimum safety factor that must be solved each time. The results show that the time required to optimize the minimum length of soil nail using neural network and genetic algorithm is greatly reduced, and it has good effect.