Effects of Likelihood Functions on Value of Information Analysis in Resource Development

摘要

Adequate information-gathering activities play a crucial role in making a rational decision under uncertainty in resource development. How adequate a particular information-gathering activity is can be quantified by using the concept of the value of information (VOI), which measures the possible increase of monetary value expected from an observation made by the information-gathering activity. In order to evaluate the VOI, one needs to model prior probability distributions for uncertain parameters, a monetary value function for each decision option, and a likelihood function representing the reliability of an observation. Among them, one often has little knowledge about a likelihood function, which is thus difficult to model. In this paper, we study the effects of likelihood functions on the VOI analysis in the context of resource development. Through numerical experiments, we show how the VOI is affected by the choice of a likelihood function. Moreover, we discuss how to include unknown bias of an observation in the definition of the VOI, and show its effect on the VOI analysis.