Search from an unknown distribution: An explicit solution (Q1196664)

This paper derives the optimal strategy of an agent who searches with recall from an unknown distribution and who holds Dirichlet beliefs. Providing his priors are accurate, this searcher will immediately accept an offer which would have terminated his search had he known for sure the distribution, but if unlucky he may eventually accept a lower wage. In the latter case, as he accumulates more experience, he is increasingly reluctant to do so. This paper also explains why some workers become discouraged.