Development of ensemble optimization approaches for managing nested multi-level environmental risks

点击翻译按钮获取中文摘要

An energy and environmental (E&E) system is composed of multiple components related to various socio-economic and environmental concerns, such as resource availability, pollution load, environmental objective, public awareness, and economic impact. Deterministic information about the system is often unavailable such that rough estimations have to be made. On the other hand, an E&E system may contain a number of subsystems at different administrative levels, demonstrating a nested hierarchical feature. Extensive interactions exist among these subsystems, implying compound complexities within national, provincial/territorial and municipal contexts. In Canada, such nested interactions exist among various socio-economic and environmental components. This is especially true with the overall demarcation of northern territories and southern provinces, where unbalanced development is accompanied with strong interdependence between them; in particular, their distinct features in socio-economic development, resources conservation and environmental protection lead to unique complexities in managing the relevant multi-level E&E issues.      The objective of this research is to develop ensemble methods (EM) for managing nested multi-level and multi-dimensional environmental risks. This entails characterization of complexities in nested E&E systems, and development of ensemble methods based on aggregation for a multitude of innovative non-deterministic optimization models. The developed approaches will then be applied to Canadian cases.     In E&E systems, uncertainties may exhibit multiple forms; moreover, among various technologies for E&E management, none dares to claim a triton among the minnows, due to uncertainties in their modeling structures, boundary conditions, and parameter arrangements. Consequently, an ensemble approach will be developed to address these complexities. Rather than relying on one model and hoping that this model is the best, the ensemble approach allows us to take a myriad of base models into account, and aggregate them to produce a more robust decision support system. It will be based on a set of innovative methods for integrated stochastic, fuzzy and interval programming and multicriteria decision analysis. Based on the developed ensemble approach, the nested vulnerabilities and risks in Canadian E&E systems will be analyzed, and desired management plans will be generated.     The proposed research represents an innovation in terms of ensemble modeling based on a set of state-of-the-art non-deterministic optimization techniques for supporting E&E management under interactive uncertainties. Application of the proposed techniques to Canadian cases will help industries and governments to improve their effectiveness in managing nested multi-dimensional E&E problems. In particular, coordinated north-south management strategies will be generated to facilitate cross-region sustainable E&E and socio-economic development.