Optimal Search from Multiple Distributions with Infinite Horizon

Abstract

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may AbstractWith infinite horizon, optimal rules for sequential search from a known distribution feature a constant reservation value that is independent of whether recall of past options is possible. We extend this result to the the case when there are multiple distributions to choose from: it is optimal to sample from the same distribution in every period and to continue searching until a constant reservation value is reached. JEL classification: D83

Designing Dynamic Research Contests

“…As Benkert, Letina, and Nöldeke (2018) argue, this model can be embedded into a Markov decision process framework. 16 This is done in the following way.…”

“…While n t (•) could potentially depend on time and state variables in complex ways, Benkert et al (2018) show that the solution to the principal's maximization problem takes a very simple form. The optimal horizon T is infinite, and there exist n F B ∈ N and…”

“…The result follows in three steps. First, by Proposition 1 in Benkert et al (2018), the first best is equivalent to a global stopping equilibrium with (i) θ F B threshold, (ii) infinite horizon, and (iii) n F B agents. Second, by Proposition 1 such an outcome can be implemented by an interim-transfer contest.…”

“…The first-best problem corresponds to the optimal search problem in Benkert et al (2018). By Proposition 1 of Benkert et al (2018) the first-best is to draw n F B observations in each period until the value of at least θ F B has been discovered.…”

“…The first-best problem corresponds to the optimal search problem in Benkert et al (2018). By Proposition 1 of Benkert et al (2018) the first-best is to draw n F B observations in each period until the value of at least θ F B has been discovered. By Proposition 1, we know that there exists an ITC which can implement the global stopping threshold θ g N with n F B N and T = ∞, thus generating the first-best surplus.…”