针对多目标优化的人工蜂群算法改进及在水文模型参数优化中的应用

项目名称： 针对多目标优化的人工蜂群算法改进及在水文模型参数优化中的应用

项目编号： No.61462058

项目类型： 地区科学基金项目

立项/批准年度： 2015

项目学科： 计算机科学学科

项目作者： 火久元

作者单位： 兰州交通大学

项目金额： 43万元

中文摘要： 鉴于目前多目标优化算法中存在的缺陷，为构建适合求解大规模、多目标问题的优化方法，本研究针对科学与工程实践中的多目标优化问题，利用随机过程理论对人工蜂群算法和多目标优化机理进行研究分析，提出一套较为完备的基于人工蜂群的多目标优化理论体系，基于该理论体系建立稳定高效的混合多目标人工蜂群智能优化算法，设计出一个灵活可配置、可独立于具体问题的多目标人工蜂群算法优化应用框架，在高性能计算环境下形成基于应用框架的水文模型参数优化方法，建立水文模型参数优化目标的综合评估机制，并将此方法应用于黑河流域SWAT、TOPMODEL等水文模型进行验证和应用前述机理、方法和技术。该研究可以更有效、更灵活地处理实际大规模多目标优化问题，对提高系统效率、促进资源合理利用具有重要意义。黑河流域中应用研究可显著提高水文模型模拟精度和参数优化的收敛速度，为干旱区内陆河流域水文模型的和水资源管理决策提供重要的方法与技术。

中文关键词： 智能计算；多目标优化；计算模型；水文模型；参数估计

英文摘要： Given the current defects that existing in the multi-objective optimization algorithm, this study is to build optimization method that suitable for solving large-scale, multi-objective optimization problem. Aimed at multi-objective optimization problems in science and engineering practice, the study is to research and analyze the Artificial Bee Colony (ABC) algorithm and multi-objective optimization mechanism by using stochastic process theory, and put forward a more comprehensive system of multi-objective optimization theory based on ABC algorithm. Based on the theoretical system to establish a stable and efficient hybrid multi-objective ABC optimization algorithm and design a flexible, configurable multi-objective ABC optimization algorithm application framework that independent of the specific problems. In the high-performance computing environments, to propose a hydrological model parameter optimization method based the application framework and establish comprehensive evaluation mechanism of parameter optimization objects in Hydrological model. This method is to be applied to the SWAT, TOPMODEL Hydrological models in Heihe River Basin for validation and application of the aforementioned mechanisms, methods and techniques. The study can be more effective, more flexibility to deal with the actual large-scale multi-objective optimization problem, and has great significance for improving the system efficiency and promoting the reasonable utilization of resources. The application research in Heihe River Basin can significantly improve the simulation precision and parameter optimization convergence rate of Hydrological model, and provide important methods and techniques for hydrological model and water resources management decision in the arid inland river basin.

英文关键词： Intelligence Computing;Multi-objective;Computing Model;Hydrological Model;Parameters Estimation