Research for Decision Theory

Syndicate content
ESTIMATING UTILITY FUNCTIONS USING PREFERENCES REVEALED UNDER UNCERTAINTY
June 1, 1974
Those who take the normative point of view of decision making under uncertainty typically suggest, sometimes explicitly, that it is sufficient to estimate utility functions on the sole basis of responses to hypothetical questions. This paper, which also takes the normative viewpoint, criticizes that approach on several grounds and proposes the alternative of combining that technique with observations of choices made under uncertainty in real life. The thesis is demonstrated using a fairly common problem in the clinical setting -- the case of severe abdominal pain in a young woman. Central to the argument is the distinction between internal inconsistency among a set of responses to abstract questions on the one hand, and disagreement between the choice deemed optimal by the theory and the choice actually revealed as most preferred by the decision maker, on the other.
Read More | Download Report
DYNAMIC PROGRAMMING MODELS OF SHORT TERM BANK RESERVE MANAGEMENT
June 1, 1974
The problem addressed here is the short-term management of a bank's reserves under uncertainty.
Read More | Download Report
AN N-STEP, 2-VARIABLE SEARCH ALGORITHM FOR THE COMPONENT PLACEMENT PROBLEM
June 1, 1974
The primary purpose of this paper is to present an n-step 2-variable placement algorithm that reliably determines near-optimal solutions to that class of problems having the mathematical structure of the quadratic assignment problem.
Read More | Download Report
A DECISION-THEORETIC APPROACH TO MEDICAL DIAGNOSIS AND TREATMENT
June 1, 1974
This paper proposes an optimal structure to the clinical decision problem.
Read More | Download Report
THE GRISLY ANALYTICS OF DEATH, DISABILITY, AND DISBURSEMENTS
June 1, 1974
In a large class of decision problems under uncertainty, death, disability, and reductions in wealth enter the analysis as crucial arguments in the utility function. Much of the literature in public health and safety programs, insurance programs, and clinical decision-making concerns itself with optimal behavior in the presence of risk in these variables. This paper sets out, primarily, to describe the assumptions which typically underlie these analyses and demonstrate why they may be too restrictive for models which deal with death, disability, and disbursement variables. An alternative scheme is proposed.
Read More | Download Report