摘要
针对基本的粒子群算法存在训练早熟等问题,调整了算法中的粒子飞行策略,将其应用于大坝变形安全监控模型中,比较了改进前后模型的效果;并建立了逐步回归统计模型,通过对比模型的预测特征参数,比较分析了模型间的预测效果。结果表明,改进的粒子群算法具有改善早熟现象的优点,且预测效果有所提高。
Aiming at the training premature problems of basic particle swarm algorithm （PSO）, the particle flight strategy in PSO is adjusted and it is applied to the model of dam safety monitoring. And the model effects are compared. Then it establishes a stepwise statistical regression model. Finally, the model prediction effect is comparative analysis by comparing the characteristic parameters of the model. The results show that the improved PSO has the advantages of improving premature and the prediction effect is raised.
出处
《水电能源科学》
北大核心
2011年第11期70-72,共3页
Water Resources and Power
基金
国家自然科学基金资助项目(50809025
50879024)
