基于改进粒子群算法的重力坝断面优化研究

Study on Cross Section Optimization of Gravity Dam Based on the Improved PSO

将粒子群算法与罚函数相结合,把非线性约束优化问题转化为无约束优化问题,解决工程上多约束优化问题。为了防止粒子群算法陷入局部最优,引入退火算法帮助粒子跳出局部最优解,从而避免粒子后期单一方向进化问题,改善粒子全局搜索能力,同时考虑到最优解一般在边界附近取得,故引入动态罚函数加强粒子对可行域边界搜索,加快最值的搜索速度。将上述方法应用于实际重力坝断面优化设计,结果表明:改进的粒子群优化算法不仅保持了良好的收敛性,而且动态罚函数还具有构造简单实用,同时退火算法减少了粒子群优化算法对大量粒子的依赖程度,验证了改进算法的有效性和可行性。

A form Combining particle swarm optimization（PSO） with the penalty function,the nonlinear constrained optimization problem was transformed into unconstrained optimization problem,and in this way the problem of multi-constrained optimization is solved.In order to avoid PSO to fall into the local optimal defects,the Anneal Arithmetic is used to help each particle to jump out of the local optimal solution,so as to avoid the single direction evolution and improved the global search ability of the particle.At the same time,considering that the optimal solution is usually obtained near the boundary,the dynamic penalty function is used to enhance the search for the feasible domain boundary and accelerate the search speed.The above method is applied to the optimization of multi-constrained gravity dam,it show that the improved PSO not only maintains a good convergence,but also the dynamic penalty function is sample to construction,and the annealing algorithm reduces the dependence of particle swarm optimization algorithm on a large number of particles,it verifies the effectiveness and feasibility of the improved algorithm.