<p>A significant majority of engineering decisions involve uncertainty and require tradeoffs between multiple attributes. Although formal methods exist to make decisions under uncertainty, often times it becomes possible to reduce or eliminate uncertainty with data collection or analysis or both. Whereas this reduction of uncertainty generally leads to better outcomes, it usually comes at an increased cost. The meta-decision of whether to reduce the uncertainty at a given cost is undertaken in the form of Value of Information (VOI) studies. &nbsp;Although previous work has considered the value of information for single-attribute cases, studies of the more challenging multiattribute case are less common.&nbsp; Multiattribute problems may have numerous uncertain variables across alternatives, and not all of them are easily measurable in money amounts. Instead, one must project the value of reduction in uncertainty on to the cost attribute, which makes the VOI analysis more complicated.&nbsp; This paper presents a simulation-based approach for calculating and interpreting the VOI in generic multiattribute decision problems and describes a simulation study that we conducted to explore how the decision problem parameters affect the VOI. A three-alternative decision problem over three attributes was considered for this study. We demonstrate that many interesting properties of value of information can be derived as a function of decision problem parameters. For example, we investigate the effect of the change in the coefficient of variation of various attributes in a decision problem. Impact of change in the mean of uncertain variables is also studied, and so is the effect of the degree of dependence between different attributes within and across alternatives.</p><p>This paper also discusses the relationship between the VOI and the gain in expected utility (utility-gain) that will result from new information.&nbsp; This relationship is not simple and yields some expected as well as counterintuitive results. For example, in certain cases the utility-gain from the new information is not correlated with the value of information. This suggests that the non-linearity of the utility functions may lead to different VoI calculation for the same utility-gain, depending on the relative position of the alternatives. Finally, this paper discusses the sensitivity of the utility from the decision as a function of decision problem parameters and derives generalizable insights from it. It is shown that the direction and magnitude of change in VoI can be predicted from these sensitivities. These results show that VoI calculation is a challenging problem that is very much contingent on the parameters of the problem. However, general guidelines can be derived using the proposed method that can aid decision making in engineering practice. A vehicle alignment problem in an automotive assembly line is presented to demonstrate the approach.</p>

Attribute-Wise Value of Information in Engineering Systems: A Simulation-Based Study

Paper Proceedings: https://asmedigitalcollection.asme.org/IMECE/IMECE2021/volume/85543

ASME 2021 International Mechanical Engineering Congress and Exposition (IMECE2021)

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.