模拟退火粒子群优化投影寻踪的渠道防渗模式评价  

Evaluation of canal seepage control plans by using simulated annealing particle swarm with projection pursuit

作　　者：董丽丽 于苗 徐淑琴[3] 

机构地区：[1]哈尔滨市水利规划设计研究院有限公司,黑龙江哈尔滨150030 [2]山东绿景水土保持工程设计咨询有限公司,山东济南250101 [3]东北农业大学水利与建筑学院,黑龙江哈尔滨150030

出　　处：《排灌机械工程学报》2016年第7期639-644,共6页Journal of Drainage and Irrigation Machinery Engineering

基　　金：黑龙江省自然科学基金资助项目(E201054)

摘　　要：为了对渠道防渗工程的方案进行优选,采用模拟退火粒子群优化投影寻踪方法对渠道防渗模式进行优化评价.优化粒子群算法,使其在退火过程中接受优解,且还以一定的概率接受差解.同时利用模拟退火粒子群算法在搜索过程中具有概率突跳的能力,有效地避免搜索过程陷入局部极小解,提高了算法的收敛速度和精度.将该模拟退火粒子群算法与投影寻踪结合,建立了模拟退火粒子群优化投影寻踪模型,将该模型应用到灌区渠道防渗模式优化评价中,选取10个指标作为评判因子,提出适合该地区的干渠渠道防渗工程优化模式,其优劣排序依次为方案Ⅳ、方案Ⅱ、方案Ⅲ、方案Ⅴ、方案Ⅰ.研究表明模拟退火粒子群优化投影寻踪的评价模型对灌区渠道防渗模式进行综合评价是切实可行的,模拟退火粒子群优化投影寻踪的评价方法在解决实际问题中展示了其优越性,并取得了良好的应用效果.Particle swarm optimization is a new technique. After simulated annealing method is intro- duced into the particle swarm optimization, the particle swarm algorithm accept good solutions, and adopt worse solutions in a certain probability. As a result, the updated algorithm avoid local minimum solutions to improve the convergence and accuracy of optimization. Further, the simulated annealing par- ticle swarm algorithm is combined with projection pursuit, then a simulated annealing particle swami op- timizing projection pursuit model is established. The model is applied to optimize irrigation canal seepage control plans based 10 indices. Finally, an optimization model is proposed to evaluate seepage controlplans for a main canal in a specific irrigation district. The results show that the proposed model is feasi- ble and the plan IV is the best, showing the proposed simulated annealing particle swarm algorithm with projection pursuit is effective and applicable.

关 键 词：渠道防渗 模拟退火 粒子群 投影寻踪 优化模式 

分 类 号：S277.9[农业科学—农业水土工程] TV93[农业科学—农业工程]