Development of a Risk-based Decision Making Framework for Hydrosystems Engineering Project Design

Problems in hydrosystems engineering management and design face two main challenges: (1) the presence of uncertainties and (2) multiple criteria/objectives in nature. Uncertainties in project design, planning, and operation induce failures rendering undesirable consequences. Involvement of several non-commensurable and often conflicting criteria creates difficulty in reaching an optimal decision. This dissertation addresses these two challenges by proposing a decision making framework according to a new decision rule based on the concept of expected opportunity loss (EOL) that can utilize pertinent uncertainty information of outcomes of alternatives to assist decision making in a highly variable environment with multiple criteria.;The EOL-based decision rule quantifies the potential risk when the chosen design/decision is incorrect. It possesses several important features that are relevant for water resources management and hydrosystem design problems: (1) circumvents the limitations of existing risk-based decision rules of using incomplete uncertain information; (2) accounts for the inter-correlations among the uncertain outcomes of different design alternatives through the mechanism of pair-wise comparison; and (3) allows feasibility test of a design alternative through comparison with decision maker's acceptable risk. By coupling with the minimax rule, the new EOL-based risk measure expands the applicability of Savage's minimax regret principle to decision making in problems involving continuous and unbounded random outcomes. Examples of evaluating flood inundation reduction projects are used to demonstrate the application of the proposed decision framework. The effect of the uncertainty level, correlation between random outcomes of alternatives, epistemic uncertainty due to sampling error, and decision maker's acceptable risk are investigated. Numerical results clearly show that the uncertainty level and correlation associated with the economic merit of hydrosystem project designs are very significant that should be not ignored.;In addition, a multi-criteria decision making (MCDM) framework is developed that explicitly considers the uncertainty features of the performance values of management alternatives under different conflicting criteria. The PROMETHEE (Preference Ranking Organization Method of Enrichment Evaluation) technique is adopted for its compatibility in decision logic and mathematical treatment with the EOL-based decision criterion involving pair-wise comparison. Numerical applications demonstrate that the EOL-based risk criterion can be conveniently incorporated in the framework of the PROMETHEE for dealing with MCDM problems under uncertainty. It opens up a new prospect to expand the prevalent practice of deterministic MCDM to probabilistic MCDM so that the uncertainty features of the performance values under different criteria are incorporated and utilized in evaluating decision alternatives in a multi-criteria problem.