Development of integrated simulation, optimization, and risk assessment methodologies for environmental management under multiple uncertainties

In this dissertation research, a set of modeling methodologies for systems simulation, optimization and risk assessment have been developed for supporting environmental management under multiple uncertainties. They include: (a) a 3-D pilot-scale physical modeling system to simulate contaminant transport and biosurfactant enhanced bioremediation processes in subsurface, (b) a 3-D inexact multiphase multi-component numerical modeling system to predict the fate and transport of contaminants under interval and fuzzy uncertainties in subsurface, (c) a hybrid decision support approach for groundwater remediation systems design, (d) hybrid optimization models to address policy analyses and systems planning issues under complexities and uncertainties, and (e) two new simulation-based risk assessment models to address multiple uncertainties associated with health impact assessment and risk management. These methodologies have been applied to a number of real-world cases in China and western Canada for environmental management and the provision of decision support in pollution control.;Reasonable results have been obtained from the 3-D pilot-scale physical modeling system and the integrated interval and fuzzy modeling system (IIFMS) based on factorial design, interval analysis, and fuzzy sets approach for predicting contaminant concentrations under hybrid uncertainties. The modeling outputs indicate system uncertainties, under a combination of information with various data-quality levels, can be effectively reflected. These simulation-based modeling systems are effective in simulating the fate and transport of contaminants with reasonably low errors, and in supporting risk assessment and decision making under uncertainty. Reasonable results have also been generated from the four optimization-based modeling methodologies including the inexact agricultural water quality management (IAWQM) model, the possibilistic stochastic water management (PSWM) model, the robust chance-constrained fuzzy possibilistic programming (RCFPP) approach, and the hybrid uncertain programming (HFICP) method with fuzzy and interval coefficients. The developed models are applied to water quality management cases for generating optimal decision schemes. They represent a unique contribution to the fields of uncertainty analysis, water quality management and policy analysis, and can effectively reflect the complex system features under various uncertainties. The obtained solutions are useful to decision makers in order to (1) gain insight into the tradeoffs among environmental, economic and system-reliability criteria, and (2) justify and/or adjust the decision schemes through the incorporation of their implicit knowledge. Reasonable results have also been generated from the integrated fuzzy interval-stochastic risk assessment (IFISRA) model and integrated fuzzy second-order stochastic risk assessment (IFSOSRA) model. The two simulation-based risk assessment models are applied to health impact assessments of groundwater quality management systems within the context of western Canada. These models incorporate subsurface modeling within a risk assessment framework, with the accuracy of modeling results being verified through comparison analysis for a series of observed and predicted data. The IFISRA model has the advantage of reflecting the complex uncertainties expressed as fuzzy, interval and stochastic parameters, as well as their hybrids such as interval-stochastic numbers in risk assessment. The IFSOSRA model provides a unique contribution to systematically quantifying variability and uncertainty presented as fuzzy, stochastic and second-order stochastic parameters in risk assessment. The distinction of variability and uncertainty in risk assessment can help decision makers better interpret the assessment outputs, and provide guidelines for better data collection. The proposed methods are especially useful in evaluating risks within a system with multiple uncertainties, and providing supports for identifying proper remedial efforts.