The three elements that form the basis of any decision are (1) The alternatives, or what we can do; (2) The information, or what we know; and (3) The preferences, or what we like. A probability distribution captures the information element of the decision and a multiattribute utility function captures the preferences in decisions with multiple objectives (attributes). The construction of a multiattribute utility function is therefore a fundamental step in decision analysis. Unfortunately, it is also a difficult task to perform in practice if we wish to capture the interactions (utility dependence relations) among the different attributes. We might, for example, have a different attitude toward placing some of our wealth at risk if we were healthy and expected to live a long time than if we were ill and expected to die soon. It is therefore important to capture this dependence relation between the two attributes if we wish to provide an accurate representation of our preferences in this decision problem. Otherwise, we may end up making the wrong decision.
The objective of this research project is to provide a general framework for constructing multiattribute utility functions that capture the utility dependence relations among the attributes in decisions with multiple objectives. The research will be validated by applying it to a variety of practical applications such as the automobile industry, global warming decisions, as well as medical decision making. The research results will be disseminated through conferences and journal papers. The algorithms will also be made available to the public through a web-based decision support system to help with decision making. In addition, the award will help provide decision training to teens at the Juvenile Detention Center in Champaign County, IL, and will also help train a group of Teens, the Peer Ambassadors group, who will learn decision skills and teach it to teens of their same age at the Juvenile Center.
From the NSF DIscoveries Section: The idea of better understanding the decision-making process could have even broader applications in the future, including within the policymaking arena--global climate change, for example--as well as in health care, design and manufacturing, education and other important fields. The work will enable better modeling of trade-offs in decisions with multiple objectives. When making a decision, there are additional factors involved, such as the frame, "meaning which decision we are really facing, as well as the logic used to make the decision, and the decision-maker, whose alternatives, information and preferences are incorporated," says Abbas. "When these elements are obtained, we can analyze any complex decision." The modelin g of trade-offs is one of those important factors. For example, someone facing medical surgery might weigh the different outcomes against their respective costs. "Suppose there were two treatment objectives: One might cost more but be less risky. The second might cost less but be riskier. How do you choose?" Abbas says. "How do we make trade-offs between different objects, such as health state and wealth, when we are uncertain about the outcomes we get? Thinking about the elements of the decision can help us make this trade-off." It is already known that people are risk-averse when it comes to money, he says. "They would rather have a sure thing than something uncertain with the same expected value," he says. "But sometimes there are other factors besides money." His research focuses on "capturing peoples' preferences over multiple attributes," says Abbas. "For example, the risk attitude over money changes with your health state. Any investments you might make, in the stock market, for example, would be different if you knew you were going to live for 10 years than if you knew you would only live for two more weeks." This change in preferences over money as health state changes is called utility dependence. "There are many things in life where a decision will change as another thing changes," he says. "It is important to capture this dependence relation between the two attributes if we wish to provide an accurate representation of our preferences in this decision problem. Otherwise, we may end up making the wrong decision." Utility dependence is a concept that appears in many decisions, both in peoples' daily lives as well as on a broader policy scale. "For example, in national security decisions, global warming decisions, hurricane events and earthquake recovery decisions, where the attributes may be money spent, lives saved, comfort level and many other natural resources," Abbas says. The proposal has also led to a workshop that brought together founders of the field of decision analysis, as well as Nobel laureates in economics with juniors in the field.
