Optimal Sampling Placement Based on Value of Information

This paper proposes an efficient method to obtain optimal observation placement based on Value of Information (Vol), which can consider not only uncertainty of estimation but risk reduction or benefit obtained by the information also. The proposed Vol-based objective function for optimal placement quantifies the risk reduction of decision making error considering the uncertainty of estimation in a Gaussian random field. The proposed method is applied to liquefaction investigation planning along an actual river embankment. The locations of the additional boring are determined by proposed Vol based method. The optimal number of additional boring locations is also evaluated for given penalty of decision-making error and observation cost.