基于参数不确定性的地下水污染治理多目标管理模型

引入随机Pareto控制排序和随机小生境技术,提出基于参数随机变化的改进小生境Pareto遗传算法,用于求解不确定性条件下地下水污染治理多目标管理模型.同时,利用顺序高斯条件模拟的蒙特卡罗方法,结合不确定性分析和风险评估,分析不同渗透系数条件点数对污染物运移结果不确定性和污染风险评价的影响.最后将该方法应用于一个考虑渗透系数为随机变量的二维地下水污染修复算例中.结果分析表明,该方法可为地下水污染治理提供变异性较小的Pareto管理策略,是一种稳定可靠的多目标随机优化方法.

A multi-objective simulation-optimization model for optimal design of groundwater remediation systems under uncertainty

The design of a robust and reliable groundwater remediation system encounters major difficulties owing to the inherent uncertainty of hydrogeological parameters. Based on the commonly used deterministic groundwater multi-objective optimization methods, a probabilistic improved niched Pareto genetic algorithm (PINPGA) is proposed for this purpose. The PINPGA uses two techniques including probabilistic Pareto domination ranking and probabilistic niche technique to find Pareto optimal solutions of groundwater remediation systems under uncertainty. Also, the performance of the proposed algorithm is evaluated through a synthetic pump-and-treat (PAT) groundwater remediation system under a random lognormal distribution of hydraulic conductivity (K) field. At first, the Sequential Gaussian Simulation (SGSIM) is used to generate conditional lnK realizations based on the sampled conditioning data acquired by the field test. Then Monte Carlo simulation is applied to address uncertainty analysis and risk assessment of contaminant transport fate associated with different numbers of conditional lnK points. Compared with the existing improved niched Pareto genetic algorithm (INPGA) with a simple averaging approach, the proposed PINPGA with a probabilistic and small sample size (as few as 5) of lnK realizations can find Pareto optimal solutions with lower variability and higher reliability, leading to a robust decision-making.