This paper addresses the problem of economic resource allocation for additional data acquisition using the value of information methodology. The basic condition to apply this method is that at decision time the available information does not describe conclusively the state of nature. Consequently, the new information might lead to a change in the decision to be made because it has the ability to modify positively the prior perception about the state of nature. The maximum value of the new information comes from how well this perception, in probabilistic terms, is modified from its original value towards the probability of conclusive or perfect information value. The new information value is calculated before acquiring it; using Bayesian theory, with decision tree methodology in conjunction with Monte Carlo simulation applied to the estimated ultimate recovery (EUR) algorithm.


Much of the previous work in estimating value of information in the petroleum industry has been deciding the use of a particular application such performing a new 3-D seismic study, running a porosity log, drilling an appraisal well, etc. The underlying principle is the comparison between the possible economic worth with information and the possible economic worth without information. The approach is correct but incomplete. Where should the economic resources be placed when a basket of information sources is available? How does this economic worth comparison behave when it is evaluated in terms of the information's usage in reserves estimates context? This study(1) was originated to find an answer to these questions.

Strategic decisions in the petroleum industry are based principally on estimated reserves. However, almost every reserve related decision is made under uncertainty because conclusive information about the reservoir is unavailable at the decision time. It is universally accepted that the uncertainty level of an estimate is related to the amount and quality of data, and hence acquisition of more data should reduce the uncertainty (2). This statement is based on Arps' work (3) on reserve estimation and uncertainty tracking.

It is less universally accepted that information does have a cost and it does or does not have a value. Costs are generally known before gathering information, i.e. from contracts, price lists, and direct negotiation with service companies. It is less intuitive to assign value to the information before gathering it.

Value of Information

In general, value of information calculation is performed using the following equation:

Equation (1) (Available in full paper)

The expected monetary value (EMV) calculation is used because there is uncertainty affecting some of the variables in a given model.

In a petroleum reserves context, the general uncertainty is a composite of the variables' particular uncertainty. An example is calculating estimated ultimate recovery (EUR) using equation (2). The techniques used so far target the identification of the most important variables and then valuing the information for those variables (4).

Equation (2) (Available in full paper)

The methodology developed in this study, goes one step further in assigning value to the variables involved in equation (2).

This content is only available via PDF.
You can access this article if you purchase or spend a download.