基于蚁群算法的复合形法及其在边坡稳定分析中的应用

The complex method based on ant colony algorithm and its application to the slope stability analysis

首先将基于排序的路径选择方法引入基本蚁群算法 ,并用之于连续变量的优化问题和边坡的最小安全系数搜索 ,结果发现对于设计变量较少的数值优化问题和简单边坡的最小安全系数搜索问题 ,该蚁群算法可以找到全局最优解或比较接近全局最优解。但对于复杂边坡的最小安全系数搜索问题 ,该蚁群算法很容易陷入局部最优。另外复合形法对于不同的初始复合形也会得到不同的最小安全系数 ,利用本文提出的基于最小海明距离的替换准则将蚁群算法得到的局部最优解替换掉初始复形中的一个顶点 ,则复合形法容易找到全局最优 ,成为一种全局搜索能力很强的优化算法。

Firstly the choosing method based on sort is imported into the basic ant colony algorithm and the new ant colony algorithm is applied to the optimization problem of the consecutive variable and the search for the minimum safety factor of the slope. The results show that the new ant colony algorithm is applicable to the few-variable numerical optimization problem and the search for the minimum safety factor of the simple slope , however the new ant colony algorithm is apt to the local minimum for the search of the minimum safety factor for the complex slope. The minimum safety factors gained by the complex method vary with the different initial multi-forms. One point of the initial multi-form is replaced with the local minimum gotten by the new ant colony algorithm using the substitution rule based on the minimum hamming distance and the complex method is apt to find the global minimum safety factor.