MULTIPERIOD SEARCH MODELS FOR AN UNKNOWN NUMBER OF VALUABLE OBJECTS

In this paper we examine multiperiod search models for cases in which the number of valuable objects is unknown. The objective is to maximize the expected total returns during the planning horizon, subject to an effort constraint. Using a Bayesian approach, we examine the model for three different priors for the number of valuable objects, and we show that the different priors (binomial, Poisson and negative binomial) lead to such conceptually different results as adaptive and nonadaptive optimal policies. The models can be applied to many areas including mineral explorations, marketing promotion activities and intelligence information.