Department of Civil Engineering, Faculty of Engineering, Architecture and Science, Ryerson University, 350 Victoria Street, Toronto, Ontario, Canada M5B 2K3.
J Hazard Mater. 2010 Apr 15;176(1-3):527-34. doi: 10.1016/j.jhazmat.2009.11.061. Epub 2009 Nov 14.
A new stochastic optimization model under modeling uncertainty (SOMUM) and parameter certainty is applied to a practical site located in western Canada. Various groundwater remediation strategies under different significance levels are obtained from the SOMUM model. The impact of modeling uncertainty (proxy-simulator residuals) on optimal remediation strategies is compared to that of parameter uncertainty (arising from physical properties). The results show that the increased remediation cost for mitigating modeling-uncertainty impact would be higher than those from models where the coefficient of variance of input parameters approximates to 40%. This provides new evidence that the modeling uncertainty in proxy-simulator residuals can hardly be ignored; there is thus a need of investigating and mitigating the impact of such uncertainties on groundwater remediation design. This work would be helpful for lowering the risk of system failure due to potential environmental-standard violation when determining optimal groundwater remediation strategies.
应用一种新的建模不确定性下的随机优化模型（SOMUM）和参数确定性，对位于加拿大西部的一个实际场地进行了研究。从 SOMUM 模型中获得了不同显著水平下的各种地下水修复策略。比较了建模不确定性（代理-模拟器残差）对最优修复策略的影响与参数不确定性（源于物理性质）的影响。结果表明，为减轻建模不确定性影响而增加的修复成本将高于那些输入参数的变异系数接近 40%的模型。这提供了新的证据，表明代理-模拟器残差中的建模不确定性几乎不能被忽略；因此，有必要研究和减轻这种不确定性对地下水修复设计的影响。这项工作有助于降低在确定最优地下水修复策略时，由于潜在的环境标准违规而导致系统故障的风险。
