Stochastic dominance: an approach to decision making under risk

Decision making under uncertainty often requires evaluating multiple options with probabilistic outcomes. Traditional approaches rely heavily on assumptions about utility functions, which may not always be available or reliable. Prior research has shown that utility-based models can be limited when preferences are unclear or incomplete. This gap motivated the development of alternative methods that require less information about decision maker preferences. Stochastic dominance offers a framework for comparing options without full knowledge of utility functions. It allows for the elimination of clearly inferior choices based on probabilistic dominance criteria. No prior work had resolved how to handle complex risk scenarios with minimal preference assumptions. This paper addresses that limitation by introducing a new decision-making tool.

The study aims to present stochastic dominance as a practical method for decision making under risk. It focuses on reducing the number of viable options a decision maker must evaluate. The approach is designed to function with limited information about individual preferences. By eliminating suboptimal choices, it simplifies the decision-making process. The method is particularly useful when utility functions are difficult to estimate or unavailable. The authors propose that this technique can be applied across various domains involving risk. Two specific scenarios are used to demonstrate its applicability. The goal is to show how stochastic dominance can guide decisions without requiring detailed preference data.

The study employs stochastic dominance as a comparative framework for evaluating options under risk. It uses probabilistic dominance criteria to identify and eliminate inferior choices. The method does not require a complete utility function from the decision maker. Instead, it relies on the relative likelihood of outcomes across options. The authors apply the technique to two real-world scenarios: building upgrades for earthquake resistance and LNG facility site selection. They compare the outcomes of each option against probabilistic thresholds. The analysis focuses on eliminating actions that are clearly dominated by others. The procedure emphasizes the reduction of choice sets rather than identifying a single optimal solution.

The application of stochastic dominance successfully eliminated certain actions as unacceptable in both scenarios. In the building upgrade example, some designs were ruled out based on their probabilistic performance. For the LNG facility site selection, the method narrowed down viable locations by comparing risk profiles. The technique did not identify a single optimal choice in either case. Instead, it provided a reduced set of options for further evaluation. The results suggest that stochastic dominance is effective in filtering out clearly inferior choices. The method requires minimal assumptions about decision maker preferences. The outcomes demonstrate its usefulness in complex risk environments where preferences are uncertain.

The authors conclude that stochastic dominance is a valuable tool for decision making under risk. It allows for the elimination of certain actions without needing detailed information about utility functions. The method is particularly useful in scenarios where preferences are incomplete or difficult to quantify. The results from the two applications support the effectiveness of the approach. The study does not claim that stochastic dominance identifies an optimal action. Instead, it emphasizes its role in reducing the decision set. The technique is proposed as a practical alternative to traditional utility-based models. The authors suggest that it can be applied in various fields involving risk assessment and choice under uncertainty.