Strategic Experimentation with Exponential Bandits.
We analyze a game of strategic experimentation with two-armed bandits whose risky arm might yield payoffs after exponentially distributed random times. Free-riding causes an inefficiently low level of experimentation in any equilibrium where the players use stationary Markovian strategies with belie...
| Main Authors: | , , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
2005
|
| _version_ | 1797065616204496896 |
|---|---|
| author | Keller, G Rady, S Cripps, M |
| author_facet | Keller, G Rady, S Cripps, M |
| author_sort | Keller, G |
| collection | OXFORD |
| description | We analyze a game of strategic experimentation with two-armed bandits whose risky arm might yield payoffs after exponentially distributed random times. Free-riding causes an inefficiently low level of experimentation in any equilibrium where the players use stationary Markovian strategies with beliefs as the state variable. We construct the unique symmetric Markovian equilibrium of the game, followed by various asymmetric ones. There is no equilibrium where all players use simple cut-off strategies. Equilibria where players switch finitely often between experimenting and free-riding all yield a similar pattern of information acquisition, greater efficiency being achieved when the players share the burden of experimentation more equitably. When players switch roles infinitely often, they can acquire an approximately efficient amount of information, but still at an inefficient rate. In terms of aggregate payoffs, all these asymmetric equilibria dominate the symmetric one wherever the latter prescribes simultaneous use of both arms. |
| first_indexed | 2024-03-06T21:31:11Z |
| format | Journal article |
| id | oxford-uuid:44bb2467-b5ad-470b-9bb7-3a9f55ca7ea9 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-06T21:31:11Z |
| publishDate | 2005 |
| record_format | dspace |
| spelling | oxford-uuid:44bb2467-b5ad-470b-9bb7-3a9f55ca7ea92022-03-26T15:03:25ZStrategic Experimentation with Exponential Bandits.Journal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:44bb2467-b5ad-470b-9bb7-3a9f55ca7ea9EnglishOxford University Research Archive - Valet2005Keller, GRady, SCripps, MWe analyze a game of strategic experimentation with two-armed bandits whose risky arm might yield payoffs after exponentially distributed random times. Free-riding causes an inefficiently low level of experimentation in any equilibrium where the players use stationary Markovian strategies with beliefs as the state variable. We construct the unique symmetric Markovian equilibrium of the game, followed by various asymmetric ones. There is no equilibrium where all players use simple cut-off strategies. Equilibria where players switch finitely often between experimenting and free-riding all yield a similar pattern of information acquisition, greater efficiency being achieved when the players share the burden of experimentation more equitably. When players switch roles infinitely often, they can acquire an approximately efficient amount of information, but still at an inefficient rate. In terms of aggregate payoffs, all these asymmetric equilibria dominate the symmetric one wherever the latter prescribes simultaneous use of both arms. |
| spellingShingle | Keller, G Rady, S Cripps, M Strategic Experimentation with Exponential Bandits. |
| title | Strategic Experimentation with Exponential Bandits. |
| title_full | Strategic Experimentation with Exponential Bandits. |
| title_fullStr | Strategic Experimentation with Exponential Bandits. |
| title_full_unstemmed | Strategic Experimentation with Exponential Bandits. |
| title_short | Strategic Experimentation with Exponential Bandits. |
| title_sort | strategic experimentation with exponential bandits |
| work_keys_str_mv | AT kellerg strategicexperimentationwithexponentialbandits AT radys strategicexperimentationwithexponentialbandits AT crippsm strategicexperimentationwithexponentialbandits |