蚁群算法在黑河上游VIC模型参数校正中的应用

参数优选是水文模型应用过程中的一项基础性工作.蚁群算法结合了分布式计算和正反馈机制,是一种较容易理解和实现的元启发式算法,已在求解复杂组合问题中展示出优异的性能.本文将蚁群算法应用于黑河上游VIC模型的参数优选中,通过与SCE-UA算法对比,探究蚁群算法在VIC模型中的适用性.经过蚁群算法优选的VIC模型在率定期(2003—2006年)和验证期(2007—2008年)的Nash效率系数分别为0.62和0.65,结果优于SCE-UA算法模拟结果.通过对蚁群算法在应用过程中的参数设定进行初步探究,结果表明:当蚂蚁数目为60,信息素蒸发系数为0.2时,蚁群算法在黑河上游水文模拟中易获得较好的率定结果.研究结果显示:蚁群算法是一种有效的VIC模型参数优选方法,适宜在其他水文模型参数优化进行推广.

Ant colony optimization in parameter calibration for the variable infiltration capacity (VIC)model in the upper Heihe River basin

Parameter calibration is a fundamental task for the application of hydrological models.Ant colony optimization (ACO)algorithm is a meta-heuristic algorithm and it shows a strong ability in tackling combinatorial problems, suitable for hydrological model calibration. In this study, ACO was applied to parameter calibration of variable infiltration capacity (VIC)model in the upper Heihe River basin,China. Shuffled complex evolution algorithm (SCE-UA)was used to test applicability of ACO.It is found that the ACO is capable of model calibration for VIC.Nash—Sutcliffe coefficient of efficiency is 0.62 in calibration period,and 0.65 in validation period,rather similar to SCE-UA results.The strategies of ACO are also discussed.Influence of the two most sensitive parameters of ACO is further investigated.The best performance of ACO is achieved when ant number is 60 and pheromone evaporation rate is 0.2.It is concluded that ACO is an effective global optimization method to calibrate large scale hydrological model.This method is also suitable for other hydrological models.