Value Of Perfect Information Research Articles

Identifying key parameters in cost-effectiveness analysis using value of information: a comparison of methods

Decisions in health care must be made, despite uncertainty about benefits, risks, and costs. Value of information analysis is a theoretically sound method to estimate the expected value of future quantitative research pertaining to an uncertain decision. If the expected value of future research does not exceed the cost of research, additional research is not justified, and decisions should be based on current evidence, despite the uncertainty. To assess the importance of individual parameters relevant to a decision, different value of information methods have been suggested. The generally recommended method assumes that the expected value of perfect knowledge concerning a parameter is estimated as the reduction in expected opportunity loss. This method, however, results in biased expected values and incorrect importance ranking of parameters. The objective of this paper is to set out the correct methods to estimate the partial expected value of perfect information and to demonstrate why the generally recommended method is incorrect conceptually and mathematically.

The value of information in a long–term hydro–thermal electrical planning model

This article presents the analysis, comparison, and application of two alternative models to the optimal long–term operation planning of an hydro–thermal power system under conditions of uncertainty. The electrical system considered comprises one large reservoir, with interannual regulation capacity, and several smaller ones. The analyzed models employ stochastic dynamic programming as the solution methodology. The fundamental problem is to decide, on every temporal stage, how much water to use for generating purposes and how much to store, in order to minimize the total thermal and shortage costs. The original version of the studied model, created originally to forecast fuel consumption, assumes that the decision regarding the water release from the main reservoir is taken knowing the future hydrologic conditions. This criterion is known as wait–and–see. On the contrary, the new versions of the model, proposed in this article, consider a here–and–now criterion. Specifically, it is assumed that the future hydrologic conditions are not known at the time of making the operational decisions. The difference between the optimal cost of the proposed models and the original model defines the value of having the information regarding future hydrologic conditions before taking any decision. This value is generally known as the expected value of perfect information.

Sensitivity Analysis to the Value of p of the l p Distance Weber Problem.

In this paper we analyze the sensitivity of the lp distance Weber problem to the value of p. We consider each power p to be in a range and not necessarily the same for each demand point. We find the set of possible optimal locations when p is in a given range. We also find the location that minimizes the expected cost, and find the Expected Value of Perfect Information. Computational results are presented.

REAL OPTIONS VALUATION AND ITS RELATIONSHIP TO BAYESIAN DECISION-MAKING METHODS

ABSTRACT It is well accepted that conventional NPV criterion fails to capture investment flexibility, and the market approach using riskless-arbitrage-pricing is ideally suited to price real options. However, when valuing complex real options, it is difficult to satisfy the restrictive assumptions required for risk-free arbitrage pricing. Using two-action linear payoff analysis, we show that when it is possible to delay and obtain additional information, an irreversible capital investment decision should be valued as an option taking into considering the value of flexibility. This option value is not based on risk-less arbitrage, but on a more fundamental concept in decision theory - the opportunity loss criterion. Our approach relates to the Quasi-Option concept of Arrow and Fisher (1974) and Henry (1974) that considers the value of gaining more information before making irreversible environment decisions. Lund (1991) provides an excellent analysis of the relationship between the Arrow and Fisher's [1] Quasi-Option Value and Black and Schole's Market based Model [4], and suggests using both ideas for valuing real options. Conrad (1980), Fisher and Hanemann (1987) and Hanemann (1989) discuss the relationship between Quasi-Option Value and Expected Value of Perfect Information (EVPI) with respect to environmental decisions. We extend the opportunity loss concept to value real options and analyze its relationship to EVPL We show that the value of a quasi-real option is equal to value of information. In the special case where a lognormal terminal distribution is assumed, we show that the EVPI is equivalent to the Black and Scholes model. We demonstrate how EVPI can be used to make investment decisions under uncertainty within an options framework. The suggested approach allows revision of option values sequentially using Bayesian methods at each decision point within the well-known decision theory framework. Financial economists have not considered the intersection between Bayesian decision framework and value of investment flexibility, that allow for a less restrictive set of assumptions.

The Value of the Value of Information

An important application of decision analysis is determining the value that information has to a decision maker. The expected value of information (EVOI) is the expected increase in the value (or decrease in the loss) associated with obtaining more information about quantities relevant to the decision process. The EVOI can be thought of as a measure of the importance of the uncertainty about a quantity in terms of the expected improvement in the decision that might be obtained from having additional information about it. Examples of EVOI quantities useful in risk management situations include the expected value of including uncertainty (EVIU), the expected value of perfect information (EVPI), and the expected value of sample information (EVSI). Value of information (VOI) analysis is useful because it makes the losses associated with decision errors explicit, balances competing probabilities and costs, helps identify the decision alternative that minimizes the expected loss, prioritizes spending on research, quantifies the value of the research to the decision maker, and provides an upper bound on what should be spent on getting information.

Evidence Propagation and Value of Evidence on Influence Diagrams

In this paper, we introduce evidence propagation operations on influence diagrams, a concept of the value of evidence to measure the impact/value of new observations/experimentation, and a concept of the value of revelation. Evidence propagation operations are critical for the computation of the value of evidence, general update and inference operations in normative expert systems that are based on the influence diagram (generalized Bayesian network) paradigm. The value of evidence allows us to compute the outcome sensitivity directly defined as the maximum difference among the values of evidence, and the value of perfect information, as the expected value of the values of evidence. We define the value of revelation as the optimal value of the values of evidence. We discuss the relationship between the value of revelation and the value of control. We also discuss implementation issues related to computation of the value of evidence and the value of perfect information.

Action-timing problem with sequential Bayesian belief revision process

We consider the problem of deciding the best action time when observations are made sequentially. Specifically we address a special type of optimal stopping problem where observations are made from state-contingent distributions and there exists uncertainty on the state. In this paper, the decision-maker's belief on state is revised sequentially based on the previous observations. By using the independence property of the observations from a given distribution, the sequential Bayesian belief revision process is represented as a simple recursive form. The methodology developed in this paper provides a new theoretical framework for addressing the uncertainty on state in the action-timing problem context. By conducting a simulation analysis, we demonstrate the value of applying Bayesian strategy which uses sequential belief revision process. In addition, we evaluate the value of perfect information to gain more insight on the effects of using Bayesian strategy in the problem.

Uncertainty, learning and international environmental policy coordination

In this paper we construct a simple model of global warming which captures a number of key features of the global warming problem: (i) environmental damages are related to the stock of greenhouse gases in the atmosphere; (ii) the global commons nature of the problem means that these are strategic interactions between the emissions policies of the governments of individual nation states; (iii) there is uncertainty about the extent of the future damages that will be incurred by each country from any given level of concentration of greenhouse gases but there is the possibility that at a future date better information about the true extent of environmental damages may become available; an important aspect of the problem is the extent to which damages in different countries may be correlated. In the first part of the paper we consider a simple model with two symmetric countries and show that the value of perfect information is an increasing function of the correlation between damages in the two countries in both the cooperative and non-cooperative equilibria. However, while the value of perfect information is always non-negative in the cooperative equilibrium, in the non-cooperative equilibrium there is a critical value of the correlation coefficient below which the value of perfect information will be negative. In the second part of the paper we construct an empirical model of global warming distinguishing between OECD and non-OECD countries and show that in the non-cooperative equilibrium the value of perfect information for OECD countries is negative when the correlation coefficient between environmental damages for OECD and non-OECD countries is negative. The implications of these results for international agreements are discussed.

Value of Traffic Assignment and Flow Prediction in Multiattribute Network Design: Framework, Issues, and Preliminary Results

The well-defined concept of value of perfect information (VOPI) was used to assess the value of improving flow prediction and the relative value of improving components of the prediction system. The concept is introduced with a simplified example of choosing whether to build a highway segment in a corridor, and then the example is extended to a network and more realistic components are incorporated. The examples are worked through to illustrate the general approach, the types of results that could be obtained, and the issues that arise. The probability distributions of the attributes—cost, time, fuel consumption, and vehicle emissions—used to summarize uncertainty in prediction are important components when calculating VOPI. An approach to modeling these distributions and the flow distributions on which they are conditioned is presented. Unlike traditional approaches, the present one recognizes uncertainty in the model itself and not only in the inputs and parameters of the model. The VOPI of arc flows is calculated for the network example and is compared with the project costs. VOPI is used to indicate that the marginal value of developing an error-free traffic assignment model would be much greater than that of developing an error-free trip distribution model in the example. The research and implementation issues raised by the example are discussed.

A NEW METHOD FOR INFLUENCE DIAGRAM EVALUATION

As influence diagrams become a popular representational tool for decision analysis, influence diagram evaluation attracts more and more research interests. In this article, we present a new, two‐phase method for influence diagram evaluation. In our method, an influence diagram is first mapped onto a decision graph and then the analysis is carried out by evaluating the decision graph. Our method is more efficient than Howard and Matheson's because, among other reasons, our method generates a much smaller decision graph for the same influence diagram. Like those most recent algorithms reported in the literature, our method also provides a clean interface between influence diagram evaluation and Bayesian net evaluation. Consequently, various well‐established algorithms for Bayesian net evaluation can be used in influence diagram evaluation. Furthermore, our method has a few unique merits. First, it takes advantage of asymmetry in influence diagrams to avoid unnecessary computation. Second, by using heuristic search techniques, it provides an explicit mechanism for using heuristic information that may be available in a domain‐specific form. These additional merits make our method more efficient than the current algorithms in general. Finally, by using decision graphs as an intermediate representation, the value of perfect information can be computed in a more efficient way.

Uncertainty in process design and operations

The paper reviews recent research developments in the area of process design and operations under uncertainty. Instead of being exhaustive, the main emphasis in the presentation is placed in linking together various components regarding the characterization and quantification of uncertainty (such as flexibility, controllability, reliability and value of perfect information) in process optimization problems in a unified theoretical way. Modelling issues and numerical techniques are discussed in detail; future research trends are also identified.

Multivariate screening procedures for quality cost minimization

We extend recent works by Tang and Tang dealing with the quality cost of a process that is screened with the use of correlated variables because the principal quality characteristic is difficult or expensive to measure. Our model includes a very flexible and general set of loss functions that can be used to model both traditional and contemporary ‘losses’ for deviation from target values or product specifications. We also give an analysis of the expected value of perfect information and sample information obtained from the screening variables. This facilitates the evaluation of the advisability of using screening variables.

The value of perfect information in a simple investment problem

The paper studies the value of obtaining perfect information in the problem of investing a given income in basic contingent claims. The class of agents for whom the value of information is monotone in the worth of the contingent claims and their riskiness are identified. Conditions which make the value of information and the willingness to pay for perfect information maximal are also given.

Potential Value of Soil Conservation Service Streamflow Outlook Based on Flexible Crop Combinations

The potential value of Soil Conservation Service (SCS) water supply outlook information was examined with a representative case farm situation in which crop enterprise combinations and irrigation practices could be adjusted in response to expected water supplies over a simulated 20-yr period (1969-1988). A programming model of a 1000-acre farm unit was used to develop optimum crop combinations and specified irrigation levels based on expected water supplies on 1 March. In this setting, overly pessimistic water supply expectations could result in foregone opportunities if more water was received than expected. Conversely, overly optimistic expectations could result in losses of abandoned acreage and unrecovered operating costs when expected water failed to materialize []

Using decision analysis for stand-level silvicultural decisions

A decision analytic approach to assessing stand-level silvicultural decisions is illustrated using planting density and spacing decisions for coastal Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco). Information from several stand projection procedures was combined into a single analysis. The value of "perfect information" was calculated as an indicator of the value of additional response information. The value of perfect information should not be used as the sole justification for gathering treatment response information because it fluctuated with the economic assumptions made and the alternatives evaluated. However, it provides a relative ranking of treatment response requirements and a means for quantifying the risk associated with silvicultural decisions. The decision analytic approach proved flexible as a means for combining growth response information from several sources into a compromise decision.

The value of perfect information in nonlinear utility theory

Wakker (1988) has recently shown that, in contrast to an expected utility maximizer, the value of information will sometimes be negative for an agent who violates the independence axiom of expected utility theory. We demonstrate, however, that the value ofperfect information will always be nonnegative if the agent satisfies a weak dominance axiom. This result thus mitigates to some degree the normative objection to nonlinear utility theory implicit in Wakker's finding.

Stochastic Sensitivity Analysis

ABSTRACT Stochastic sensitivity analysis explicitly considers the uncertainty common in engineering economy. Like simulation, the variables probability distributions are the input. The output, like deterministic sensitivity analyses, relates changing variables to changes in the model's outcome (present worth). However, the x —axis is now probabilistic so that break-even probabilities are r-axis distances. Areas represent conditional expected values and expected values of perfect information. This solid intuitive basis contrasts with a uniform pdf assumption that underlies typical sensitivity graphs. The probability metric permits relative sensitivity analysis of variables that are measured in different units. Spreadsheets easily support this method.

Techniques for Assessing the Economic Contributions of Forest Soils Information in Timberland Management

Abstract Forest soils information can be very useful in helping forest landowners achieve specific objectives. Major factors that should be considered in the evaluation of forest soils information include the management decision level at which it is utilized, the frequency of use, and the type of control process in which the information functions. In principle, information essential in formulating strategic plans is normally of greater value than information that is needed to control operational or "on the ground" activities, however significant exceptions occur. Because of the long time frame needed to adequately evaluate growth responses on lands actually treated, soils information generally functions through control processes that are predictive in nature. A marginal risk-benefit approach for discriminating between alternative management options can be used to identify useful information. Several techniques can be used in determining the potential utility of soils information, such as the mean-variance rule, stochastic efficiency analysis, and the expected value of perfect information. South. J. Appl. For. 12(3):147-152.

Dairy Cow Culling Decision. 3. Risk of Culling on Predicted Income (An Application of Bayes Criterion)

Comparison of cows on their potential for income production or cash flow over multiple calving intervals has been utilized for culling. The reliability of income projections several years in the future may be questioned when the economic environment is uncertain. The problem would be apparent when cull candidates differ in age since a mature cow would be selected for immediate productivity whereas the immature cow would be selected for her long-term potential. When an uncertain change in the economic environment (e.g., milk price) results in a difference in ranking of two cull candidates, the Bayes criterion of selecting the animal with the highest expected value for income can be utilized. A sensitivity analysis can then determine the dependence of the decision on variation in milk price or other states of nature. The value of perfect information would assess the financial risk associated with the decision. In an example analyzed with a probable 5% decline in milk prices, information that would reduce the variability in predicting animal performance had more effect on the payoff of the culling decision than more accurate predictions of future milk prices.

Partial Means in the Economic Risk Analysis of Projects

Abstract Partial means constitute a new measure of economic risk that unifies and subsumes many less general measures. The partial mean of a random variable is its expected value conditioned on its falling within a given range. For a random variable such as the present worth of a return with uncertain timings, partial means are directly usable to discretize a continuum of outcomes in order to build a decision tree. More importantly, partial means can be used as building blocks in computing a variety of established risk measures - loss integrals, Fishbum's risk measure, the expected value of perfect information - thus unifying these measures and elucidating their interrelationships. In this paper we define partial means and related measures, illustrate their characteristics through numerical examples, and show their relationships to several less general risk measures. Tables of partial mean properties are presented for various probability distributions, to facilitate computations and comparisons among comp...