Granularity Helps Explain Seemingly Irrational Features of Human Decision Making

Abstract

Starting from well-known studies by Kahmenan and Tarsky, researchers have found many examples when our decision making -and our decision making -seem to be irrational. In this chapter, we show that this seemingly irrational decision making can be explained if we take into account that human abilities to process information are limited; as a result, instead of the exact values of different quantities, we operate with granules that contain these values. On several examples, we show that optimization under such granularity restriction indeed leads to observed human decision making. Thus, granularity helps explain seemingly irrational human decision making. Known Examples of Seemingly Irrational Human Decision Making: Formulation of the ProblemIn the ideal world, people should make perfect decisions. In many real-life situations, we know what is best for us, and we know the exact consequences of each of our actions. In this case, a rational person should select an action that leads to the best possible outcome. This assumption underlies basic (idealized) economic models: in these models, our decision making may hurt others but every person is interesting in selecting a decision which is the best for him/herself.In the real world, people's decisions are not perfect. In the perfect world, people should make perfect decisions. It is well known, however, that our world is not perfect, and that many people make decisions which are not in their own best interests. People eat unhealthy food, fail to exercise, get drunk, smoke, take drugs, gamble, and do many other things which -as they perfectly know -are bad for their health and bad for their wallets.Such imperfect decisions can still be described in optimization terms. People engage in all kinds of unhealthy and asocial decision making because they get a lot of positive emotions from this engagement. A drug addict may lose his money, his family, his job, his health -but he gets so much pleasure from his drugs that he cannot stop. A gambler may lose all his money, but the pleasure of gambling is so high that he continues gambling (and losing) money until no money is left.These examples of bad decisions are bad from the viewpoint of a person's health or wealth or social status. In all these examples, people clearly know what they want -e.g., more pleasure from drugs or from gambling -and they select a decision which is the "best" from this viewpoint.On top of this well-known abnormal decision making, there are also many examples when a seemingly rational decision making is actually irrational. It is well known that people make 1

How to transform partial order between degrees into numerical values

“…• For the case of linearly ordered set, a similar idea was used in [6], [7], [8], [9] to explain a seemingly irrational character of human decision making under uncertaintyas described, e.g., in [4].…”

Multi-attribute decision-making based on the SPIFGIA operators

“…Compared with the ''crisp'' properties that are either true or false, to describe the properties H that decision makers are not completely sure whether a fixed value is big or small, u H ðxÞ is assigned to each alternative to indicate the degree which it satisfies this property (Lorkowski and Kreinovich 2015): (1) the degree u H ðxÞ ¼ 1 indicates that the decision makers are completely sure that the value x satisfies the property H; (2) u H ðxÞ ¼ 0 indicates that the decision makers are completely sure that the value x does not meet H; (3) if 0\u H ðxÞ\1, then the decision makers are not completely sure that the value x satisfies the property H or not. By requesting decision makers to mark their preference degrees of x meets the property H based on a Likert scale, Lorkowski and Kreinovich (2015) used granules to draw forth u H ðxÞ from experts. For example, given a scale from 0 to n, the decision makers can choose their preferences from the labels of 0, 1,…,n, if the decision maker choose m, then u H ðxÞ will be m/n.…”

“…For example, given a scale from 0 to n, the decision makers can choose their preferences from the labels of 0, 1,…,n, if the decision maker choose m, then u H ðxÞ will be m/n. Lorkowski and Kreinovich (2015) also stated that the decision makers can choose their description of uncertainty based on a larger value n if a more detailed scale is needed and given.…”

“…Therefore, with the methods in Pedrycz and Chen (2015), Lorkowski and Kreinovich (2015) and Naim and Hagras (2015), i.e., Granular Computing, we can construct the IFNs in this paper. For example, experts marks her confidence by left membership value 6 and right membership value 8 on a scale from 1 to 10 to indicate that the degrees x 1 satisfy the property H 1 , and thus, A 11 in the numerical example is…”

Towards the most robust way of assigning numerical degrees to ordered labels, with possible applications to dark matter and dark energy

“…It turns out that this approach leads exactly to the same tuple r i = i n + 1 [2], [3], [4], [6], [7], [8], [9], [10], which makes us more confident that the above robustness approach makes sense.…”