Utility Theory

Utility Theory

https://www.cs.ubc.ca/~kevinlb/teaching/cs532l%20-%202013-14/Lectures/Utility%20Theory.pdf
A utility function is a real-valued function that indicates how much agents like an outcome.
In the presence of uncertainty, rational agents act to maximize their expected utility.
Utility is a foundational concept in game theory.
But it is a nontrivial claim:
1 Why should we believe that an agent's preferences can be adequately represented by a single number?
2 Why should agents maximize expectations rather than some other criterion?
Von Neumann and Morgenstern's theorem shows why (and when!) these are true.
It is also a good example of some common elements in game theory (and economics):
Behaving \as-if"
Axiomatic characterization

https://en.wikipedia.org/wiki/Von_Neumann%E2%80%93Morgenstern_utility_theorem
https://en.wikipedia.org/wiki/Utility
http://static.luiss.it/hey/microeconomia/book/Ch21.pdf
https://www.journals.uchicago.edu/doi/abs/10.1086/257028?journalCode=jpe


https://www.researchgate.net/publication/24099287_Assessment_of_Attribute_Importances_and_Consumer_Utility_Functions_Von_Neumann-Morgenstern_Theory_Applied_to_Consumer_Behavior

CONCLUSION

This initial application of vN-M utility theory indicates that consumer measurement is feasible, psychological attributes can be included, and empirical results were equal to or better than two competing approaches. These are encouraging results because this implies vN-M utility theory's attractive features of modeling risk, indifference measurement, and identification of practical form have potential for application in consumer research.

Von Neumann-Morgenstern utility theory can be a valuable tool for understanding and predicting consumer behavior. It can be most effective if: (1) risk aversion and interaction phenomena are deemed to be important in the consumer's behavior, (2) a sufficient budget is available for the personal interviews, (3) individual utility parameters are important to the research or managerial question, and (4) consumers are well educated. It is particularly effective if the number of decision-makers is small and the choice decision large. For example, purchase of large computers, aircraft, automated machine tools, or other industrial products might be good applications. Other useful examples might be consumer durables, such as washer / dryers or automobiles. In services it might be applicable to health care, college selection, and career selection.

Several future areas of research are appropriate. Our application acknowledges measurement error, but does not explicitly include it in parameter estimation. Research is needed to allow degrees of freedom and to develop distributional assumptions for parameter estimation. The utility and conjoint axioms appear compatible, but research is needed to develop a common set of consistent axioms. Needed also are statistical tests of assumptions so that confidence limits can be set for the repeated lottery and tradeoff questions used in testing utility and preferential independence.

Another topic is the development of more efficient measurement techniques, thus allowing some combination of more complexity, more attributes, or more assumption testing. A final need is to develop simpler measurement methods. Our sample was MIT students. Von Neumann-Morgenstern utility theory requires further testing to determine whether an average respondent could accurately answer lottery questions, even with careful training.