Oftentimes contractors have to make a decision whether to perform certain ground exploration work for additional geological data. Decision analysis techniques are usually used and help contractors make the decision. In the analysis, results are not determined by the cost of the ground exploration only but also highly depending on how reliable the exploration results are. Due to insufficient information, the analysis usually relies on judgments and experience from engineers, and the reliability or probabilities are often given imprecisely such as intervals. As a result, imprecision is involved in probability evaluations. On the other hand, current methodologies of decision analysis usually do not consider imprecision in probability evaluation. Designers have to pick a single value instead of the probability interval as the input for the decision analysis. This paper presents algorithms how to carry out the analysis when probabilities are evaluated imprecisely. It starts from an example about choosing the construction method for a tunnel to be constructed in an uncertain ground. Then the paper discusses whether to perform additional ground exploration works for more geological/geotechnical information and how much should be paid for that. A case history of a rock tunnel through various ground conditions is revisited. Decision analysis with imprecise probabilities is performed to determine the optimal exploration plan and the value of the results obtained from additional ground exploration.
