Read e-book What Every Engineer Should Know About Decision Making Under Uncertainty

Free download. Book file PDF easily for everyone and every device. You can download and read online What Every Engineer Should Know About Decision Making Under Uncertainty file PDF Book only if you are registered here. And also you can download or read online all Book PDF file that related with What Every Engineer Should Know About Decision Making Under Uncertainty book. Happy reading What Every Engineer Should Know About Decision Making Under Uncertainty Bookeveryone. Download file Free Book PDF What Every Engineer Should Know About Decision Making Under Uncertainty at Complete PDF Library. This Book have some digital formats such us :paperbook, ebook, kindle, epub, fb2 and another formats. Here is The CompletePDF Book Library. It's free to register here to get Book file PDF What Every Engineer Should Know About Decision Making Under Uncertainty Pocket Guide.
What Every Engineer Should Know About Decision Making Under Uncertainty [ John X. Wang] on anechousmul.tk *FREE* shipping on qualifying offers. Covering.
Table of contents

Subjects predicted the likelihood that the unknown person would show a specific behavior, e. However, in most conditions participants only paid attention to the target category real estate agent when making predictions, ignoring relevant information from the secondary category. Similar results have been found in studies using artificial categories. The current project extends this category-based inference research to decision making.

We had two questions. First, does decision making lead to a more rational use of multiple categories when categorization is uncertain? Second, do ideas from the decision making literature provide new ideas on how to influence thinking in these uncertain categorization situations? In particular, can we use framing effects, which often have large influences in decision making, to help overcome the reliance on single categories?

Category-based inferences are likely related to category-based decisions, as a prediction often serves as a reason for the decision. Choosing actions is perhaps one step more complex than predicting a property, because the action must be predicated upon the expected properties of the uncertain object.

Thus, perhaps the results in decision making when categorizations are uncertain will show the same single-category focus as in category-based inferences. However, whereas category-based inferences ask people to make some prediction about the future, they do not point to any consequences about the decision. Because decision making involves selecting an action, people may be more careful in thinking through possible consequences. Thus, the first goal of extending the research to decision making is not simply to generalize a body of research to another related domain, but rather to ask whether this new domain, because it focuses on taking an action, might lead to a very different result.

Mynatt et al.

What Every Engineer Should Know About decision making under uncertainty

For example, they provided subjects with some evidence about a specific car and then selected more evidence in order to either say which of two kinds of car it was inference or which of the two cars they would buy action. But when making a decision about which car to buy in the action task, people selected information about both categories that allowed them to compare them. This result suggests that the focus on single categories found in our category-based induction task may not be found in a decision-making task.

However, as we shall discuss in the General Discussion, Mynatt et al. The second goal was to use the extension to decision making to suggest additional manipulations that might promote more normative judgments, something that has proven difficult to devise with category-based inferences. In particular, we asked how framing of choice information influences how category information is used in decision making under uncertainty.

Research on decision making has shown that the way choices are presented can have enormous effects on decisions, for example, loss vs. Our goal was to attempt to discover whether use of multiple categories could be encouraged by the organization of information. If one form leads to more normative choices than another, this information could be used in designing decision frames when people must choose between uncertain options Thaler and Sunstein, To examine if framing of category information could promote multiple-category use, we presented information in two ways.

In one way of presenting the information, the outcomes were organized by the category the category frame ; the possible choices and their associated outcomes were listed underneath each category. In the other organization, the information was listed by choice the choice frame. In the category frame, all the choices for that category are listed together. It seems very easy to focus on the information for one category, as information about the possible categories was physically separated.

In contrast, in the choice frame, this was not the case: Information about the two categories was presented together for each choice. Since choices were shared across categories, this might encourage people to combine the information across categories, as described below.

Our Bestsellers

Past work on category-based induction used displays analogous to the category frame; to our knowledge, nothing like the choice frame has been tested in that paradigm. In many real-life situations, the use of categories is much more complex than in our paradigm. Fischoff and Beyth-Marom provided a classic analysis of Bayesian hypothesis testing in which they broke down the Bayes' rule into component psychological processes that were required to put it into effect e. Our task eliminates many of these problems.


  • Activated Carbon!
  • Marxism and moral concepts?
  • The Hidden Power and Other Papers Upon Mental Science [with Biographical Introduction].
  • Hazardous Materials Compliance for Public Research Organizations: A Case Study, Second Edition.
  • Engineering Decision Making & Risk Management;
  • Hauptnavigation;

For example, we specify the relevant hypotheses categories , provide their base rates priors , give all the information about their properties, and so on. Information search is almost trivial in our task, and memory requirements are minimal. The most relevant issue in their taxonomy is probably assessing the likelihood ratio: Comparing the evidence for one hypothesis to the evidence for another. As they point out p. Thus, nondiagnostic evidence may be treated as useful, if it is associated to the target hypothesis as in Doherty et al. This is quite similar to what we have found in the category-based induction case, where evidence is derived from a target category only, even when it is not certain.

Other work has emphasized that decision making may not take into account all hypotheses that might be considered. Dougherty et al. Working memory provides a limit on the number of hypotheses that are considered, a number that can be reduced with task load and other manipulations.

Account Options

We have also considered the issue of working memory in the induction context e. It seems likely that the limit on hypothesis generation in the real-world context that Dougherty et al. Indeed, the limits on the numbers of hypotheses they consider p. Another relevant issue concerns the way in which the likelihood of the different categories is communicated. In much of our previous work, the categories were visually displayed, and subjects had to extract the base rates on their own.


  • Engineering Decision Making.
  • Book Description?
  • The Operation and Regulation of Financial Markets!
  • Advances In Biochemical Engineering - Biotechnology - Modern Adva.
  • Military Uniforms?

In the present design, which did not use perceptual features, likelihood is communicated verbally, through percentages e. Use of such probabilities may make it more difficult for people to reason correctly about the categories and their properties, as illustrated by research on likelihoods transmitted by probabilities, frequencies, or direct experience e. Furthermore, visual presentations such as those used in our past work may be especially helpful in promoting Bayesian reasoning Arkes and Gaissmaier, Thus, it is possible that people will find it more difficult to attend to and reason about the category probabilities in the present experiment.

We explicitly addressed this possibility in Experiment 2 by making sure that subjects had encoded and remembered those probabilities. The first goal of Experiment 1 was to investigate how people use category information when making a decision under uncertain categorization. To do this, we presented subjects with short vignettes that involved a situation of ambiguous categorization. Subjects had to decide what the best choice for the main character of each story was.

Decision Analysis 1: Maximax, Maximin, Minimax Regret

For example, one vignette described a situation in which a patient had to choose which medication to take for a rash with an ambiguous diagnosis categorization , another described a professor who had to decide which book would be most useful for a student but could not remember whether the student was a business or science major.

See Table 1 for an example vignette. After making a decision, subjects answered one comprehension question for each vignette. This was to ensure that subjects had carefully read the information in the story.

co.organiccrap.com/69286.php

ISBN 13: 9780824708085

Following the design of Murphy and Ross b , we constructed the stories such that we could determine whether subjects were basing decisions on multiple categories or on only a single category. One choice was dominant if subjects considered only the most likely or target category, and another choice would be the dominant choice if subjects also considered the less likely or alternative category in addition to the target category. However, if paying attention to both categories, the best option was the Jones book.

These calculations follow the general Bayesian rule that the prediction of each category should be weighted by its likelihood e. However, the predictions do not depend on such weighting, which could be computationally difficult. For example, if people give equal weights to the categories, then they should also choose the Jones book, because it has moderately high success rates in both categories i. These types of simple averaging strategies will, of course, not always be successful see McKenzie, for analysis of the performance of various averaging strategies vs.

Bayesian ones. For example, if people arbitrarily give weights that differ greatly from the stated probabilities, the predictions no longer hold. However, in past research in which we have asked subjects to provide their own probabilities for the categories, they are generally quite reasonable e. The second goal of Experiment 1 was to examine how the framing of choice information influences subjects' propensity to use multiple categories in decision making.

To examine this issue, we framed the provided information in two ways, by category or by choice. The above calculations do not depend on how the various probabilities are presented.

Decision Making under Uncertainty Networks - EPSRC website

As Table 1 suggests, the category frame might encourage subjects to focus on the most likely category and therefore not even notice that one of its associated choices is not useful to the alternative category. In contrast, the choice frame presents information about a given choice together, putting information about the usefulness of that choice for all possible categories together, thereby possibly encouraging people to use multiple categories. In order to only pay attention to one category in this frame, subjects would have to selectively ignore a given category's information separately for each header.