‘Can You Fix It?’ Using Variance-Based Sensitivity Analysis to Reduce the Input Space of an Agent-Based Model of Land Use Change

‘Can You Fix It?’ Using Variance-Based Sensitivity Analysis to Reduce the Input Space of an Agent-Based Model of Land Use Change
复制标题

“你能解决这个问题吗?”使用基于方差的敏感性分析来减少基于主体的土地利用变化模型的输入空间

DOI:
--
复制
发表时间:
2018
期刊:
影响因子:
--
通讯作者:
Arika Ligmann
Arika Ligmann
中科院分区:
--
文献类型:
--
作者:
Arika Ligmann

文献摘要

参考文献

被引文献

相似文献

不断增长的基于主体建模(ABM)的知识表明,统计和系统不确定性是开发简约模型的最具挑战性的障碍。为了降低 ABM 维数,采用了综合不确定性分析 (UA) 和敏感性分析 (SA),其中使用蒙特卡洛模拟通过模型传播输入不确定性,从而产生输出的概率分布 (PD)。使用方差(一种简单而简洁的结果变异性衡量标准)进一步总结 PD。使用分解将方差分配给模型输入,以便量化其中哪些输入以及在多大程度上影响 ABM 结果的可变性。计算出的灵敏度指数代表每个输入对输出方差的贡献分数。为了简化模型,可以将低灵敏度值的输入设置为常量(例如平均值),从而有效地减少模型输入空间。
The growing body of knowledge on agent-based modeling (ABM) points to statistical and systematic uncertainty as the most challenging obstacles to developing parsimonious models. To decrease ABM dimensionality, a comprehensive uncertainty analysis (UA) and sensitivity analysis (SA) are employed, where input uncertainty is propagated through the model using Monte Carlo simulations, resulting in probability distributions (PDs) of outputs. The PDs are further summarized using variance—a simple yet succinct measure of result variability. The variance is apportioned to model inputs using decomposition, in order to quantify which of them and to what extent affect the variability of ABM results. The calculated sensitivity indices represent fractional contributions of each input to output variance. To simplify the model, inputs with low sensitivity values can be set to constants (e.g. mean), effectively decreasing model input space.
DOI: 10.1016/j.cpc.2011.12.020
发表时间: 2012-04-01
影响因子: 6.3
作者:
Kucherenko, S.;Tarantola, S.;Annoni, P.
通讯作者: Annoni, P.