智能算法反演土壤水分特征曲线参数的应用

Application of Intelligent Algorithms for Parametric Inversion of Soil Water Characteristic Curves

确定土壤水分曲线参数向来是讨论土壤水分和盐分运移的重要前提。用传统方法求非线性优化问题时,难免会出现不易收敛、费时、容易找到错误解等不利局面。现代发展起来的智能优化算法,很好地解决了这些不足。基于模仿生物生理特点和生物适应属性而开发的萤火虫算法、粒子群算法等仿生算法、基于条件概率的层次贝叶斯方法和基于同化数据原理的集合卡尔曼滤波算法都具有很强的优化能力和寻优效率。将这些新算法应用在反演土壤水分曲线参数的结果表明,智能算法完全可以用于求解该问题,而且以较快速度收敛,迅速找到最优解。从这些算法应用和效果来看,智能算法在土壤水分参数反演这一领域有广泛的应用前景。

Solving the inverse problem of soil water characteristic curves is an important part for the investigation of water and salt movement in unsaturated soil. Traditional soil moisture inverse solutions （gradient-based nonlinear optimization problems） generally have three disadvantages, such as slow convergence rate, low accuracy, and frequent occurrence of local minimum solutions. The bio-inspired optimization algorithms, such as firefly algorithm and particle swarm optimization which are developed by mimicking the animal behaviors and evolution characteristics, Hierarchical Bayesian Method which is based on conditional probability and Ensemble Kalman Filter which is based on sequential data assimilation method, all have strong capacity to search the global optimal solutions with good efficiency. The application of the new algorithms for solving the parameters inverse problem of soil water characteristic curves indicates that the new algorithms can discover globally optimal solutions with larger probability and faster convergence. According to the algorithm theories and their application effects, the intelligent algorithms have broad prospects for solving the parameters inverse problem of soil water characteristic curves