Abstract
This paper proposes a method for optimal sampling planning, i.e., the number and placement of additional sampling for land with arbitrary shape and with existing sampling data, under the assumption of Gaussian random field with respect to a characteristic parameter. A set of optimal locations for additional sampling are evaluated as a solution of optimization problem, in which its objective function is Value of Information (VoI), and the optimization method is Particle Swarm Optimization. Optimal number of sampling is also evaluated by total cost, i.e., sum of observation cost and VoI. The balance of penalties and observation cost determines the optimal number of additional sampling.
