Chen Lei, Wei Guoyuan, Shen Zhenyao
State Key Laboratory of Water Environment, School of Environment, Beijing Normal University, Beijing 100875, P.R. China.
Sci Rep. 2015 Oct 21;5:15393. doi: 10.1038/srep15393.
To solve computationally intensive and technically complex control of nonpoint source pollution, the traditional genetic algorithm was modified into an auto-adaptive pattern, and a new framework was proposed by integrating this new algorithm with a watershed model and an economic module. Although conceptually simple and comprehensive, the proposed algorithm would search automatically for those Pareto-optimality solutions without a complex calibration of optimization parameters. The model was applied in a case study in a typical watershed of the Three Gorges Reservoir area, China. The results indicated that the evolutionary process of optimization was improved due to the incorporation of auto-adaptive parameters. In addition, the proposed algorithm outperformed the state-of-the-art existing algorithms in terms of convergence ability and computational efficiency. At the same cost level, solutions with greater pollutant reductions could be identified. From a scientific viewpoint, the proposed algorithm could be extended to other watersheds to provide cost-effective configurations of BMPs.
为解决非点源污染计算量大且技术复杂的控制问题，将传统遗传算法改进为自适应模式，并通过将这种新算法与流域模型和经济模块相结合，提出了一个新框架。尽管该算法在概念上简单且全面，但无需对优化参数进行复杂校准即可自动搜索那些帕累托最优解。该模型应用于中国三峡库区典型流域的案例研究。结果表明，由于纳入了自适应参数，优化的进化过程得到了改善。此外，该算法在收敛能力和计算效率方面优于现有最先进的算法。在相同成本水平下，可以识别出污染物削减量更大的解决方案。从科学角度来看，该算法可扩展到其他流域，以提供具有成本效益的最佳管理措施配置。
