# POMDP Policy Optimization on Random POMDP Source: [POMDP Policy Optimization on Random POMDP]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:48:23.471Z Evaluation of policy optimization algorithms applied to the Random Partially Observable Markov Decision Process (POMDP) environment, typically measured using the L^π metric. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | POMDP Policy Optimization | | Dataset (including any stated split) | Random POMDP | | Metric | L^pi | | Unit | SCALAR | | Direction | Lower is better | | Results loaded | 31 of 31 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [L^pi](): Lower is better; unit: SCALAR. ## Recorded results | Method | L^pi | Result date | Report paper | | --- | --- | --- | --- | | Cohen and Parmentier (2023) | 3.1284 | Not recorded | 693f96136cfd488761a81bbd | | Memoryless Policy Iteration | 3.2125 | Not recorded | 693f96136cfd488761a81bbd | | Policy gradient | 3.2125 | Not recorded | 693f96136cfd488761a81bbd | | Exhaustive search | 3.2125 | Not recorded | 693f96136cfd488761a81bbd | | Cohen and Parmentier (2023) | 3.2125 | Not recorded | 693f96136cfd488761a81bbd | | Müller and Montufar (2022) | 3.2125 | Not recorded | 693f96136cfd488761a81bbd | | Müller and Montufar (2022) | 3.2133 | Not recorded | 693f96136cfd488761a81bbd | | Memoryless Policy Iteration | 3.214 | Not recorded | 693f96136cfd488761a81bbd | | Policy gradient | 3.214 | Not recorded | 693f96136cfd488761a81bbd | | Exhaustive search | 3.214 | Not recorded | 693f96136cfd488761a81bbd | | Müller and Montufar (2022) | 3.3865 | Not recorded | 693f96136cfd488761a81bbd | | Memoryless Policy Iteration | 3.3868 | Not recorded | 693f96136cfd488761a81bbd | | Policy gradient | 3.3868 | Not recorded | 693f96136cfd488761a81bbd | | Müller and Montufar (2022) | 3.4866 | Not recorded | 693f96136cfd488761a81bbd | | Memoryless Policy Iteration | 3.492 | Not recorded | 693f96136cfd488761a81bbd | | Policy gradient | 3.492 | Not recorded | 693f96136cfd488761a81bbd | | Müller and Montufar (2022) | 3.5745 | Not recorded | 693f96136cfd488761a81bbd | | Memoryless Policy Iteration | 3.5749 | Not recorded | 693f96136cfd488761a81bbd | | Policy gradient | 3.5749 | Not recorded | 693f96136cfd488761a81bbd | | Müller and Montufar (2022) | 3.6054 | Not recorded | 693f96136cfd488761a81bbd | | Policy gradient | 3.6059 | Not recorded | 693f96136cfd488761a81bbd | | Memoryless Policy Iteration | 3.606 | Not recorded | 693f96136cfd488761a81bbd | | Memoryless Policy Iteration | 3.6338 | Not recorded | 693f96136cfd488761a81bbd | | Policy gradient | 3.6338 | Not recorded | 693f96136cfd488761a81bbd | | Müller and Montufar (2022) | 3.6338 | Not recorded | 693f96136cfd488761a81bbd | | Müller and Montufar (2022) | 3.636 | Not recorded | 693f96136cfd488761a81bbd | | Müller and Montufar (2022) | 3.6364 | Not recorded | 693f96136cfd488761a81bbd | | Memoryless Policy Iteration | 3.6378 | Not recorded | 693f96136cfd488761a81bbd | | Policy gradient | 3.6378 | Not recorded | 693f96136cfd488761a81bbd | | Memoryless Policy Iteration | 3.6392 | Not recorded | 693f96136cfd488761a81bbd | | Policy gradient | 3.6392 | Not recorded | 693f96136cfd488761a81bbd | ### Cohen and Parmentier (2023) L^pi: 3.1284. Result ID: 69996cda5698665ec4da9b5a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"3"}. This record does not assert independent reproduction. ### Memoryless Policy Iteration L^pi: 3.2125. Result ID: 69996cda5698665ec4da9b45. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"2"}. This record does not assert independent reproduction. ### Policy gradient L^pi: 3.2125. Result ID: 69996cda5698665ec4da9b4e. [Reporting paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"2"}. This record does not assert independent reproduction. ### Exhaustive search L^pi: 3.2125. Result ID: 69996cda5698665ec4da9b57. [Reporting paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"2"}. This record does not assert independent reproduction. ### Cohen and Parmentier (2023) L^pi: 3.2125. Result ID: 69996cda5698665ec4da9b59. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"2"}. This record does not assert independent reproduction. ### Müller and Montufar (2022) L^pi: 3.2125. Result ID: 69996cda5698665ec4da9b5b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"2"}. This record does not assert independent reproduction. ### Müller and Montufar (2022) L^pi: 3.2133. Result ID: 69996cda5698665ec4da9b5c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"3"}. This record does not assert independent reproduction. ### Memoryless Policy Iteration L^pi: 3.214. Result ID: 69996cda5698665ec4da9b46. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"3"}. This record does not assert independent reproduction. ### Policy gradient L^pi: 3.214. Result ID: 69996cda5698665ec4da9b4f. [Reporting paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"3"}. This record does not assert independent reproduction. ### Exhaustive search L^pi: 3.214. Result ID: 69996cda5698665ec4da9b58. [Reporting paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"3"}. This record does not assert independent reproduction. ### Müller and Montufar (2022) L^pi: 3.3865. Result ID: 69996cda5698665ec4da9b5d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"4"}. This record does not assert independent reproduction. ### Memoryless Policy Iteration L^pi: 3.3868. Result ID: 69996cda5698665ec4da9b47. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"4"}. This record does not assert independent reproduction. ### Policy gradient L^pi: 3.3868. Result ID: 69996cda5698665ec4da9b50. [Reporting paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"4"}. This record does not assert independent reproduction. ### Müller and Montufar (2022) L^pi: 3.4866. Result ID: 69996cda5698665ec4da9b5e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"5"}. This record does not assert independent reproduction. ### Memoryless Policy Iteration L^pi: 3.492. Result ID: 69996cda5698665ec4da9b48. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"5"}. This record does not assert independent reproduction. ### Policy gradient L^pi: 3.492. Result ID: 69996cda5698665ec4da9b51. [Reporting paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"5"}. This record does not assert independent reproduction. ### Müller and Montufar (2022) L^pi: 3.5745. Result ID: 69996cda5698665ec4da9b60. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"7"}. This record does not assert independent reproduction. ### Memoryless Policy Iteration L^pi: 3.5749. Result ID: 69996cda5698665ec4da9b4a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"7"}. This record does not assert independent reproduction. ### Policy gradient L^pi: 3.5749. Result ID: 69996cda5698665ec4da9b53. [Reporting paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"7"}. This record does not assert independent reproduction. ### Müller and Montufar (2022) L^pi: 3.6054. Result ID: 69996cda5698665ec4da9b5f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"6"}. This record does not assert independent reproduction. ### Policy gradient L^pi: 3.6059. Result ID: 69996cda5698665ec4da9b52. [Reporting paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"6"}. This record does not assert independent reproduction. ### Memoryless Policy Iteration L^pi: 3.606. Result ID: 69996cda5698665ec4da9b49. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"6"}. This record does not assert independent reproduction. ### Memoryless Policy Iteration L^pi: 3.6338. Result ID: 69996cda5698665ec4da9b4b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"8"}. This record does not assert independent reproduction. ### Policy gradient L^pi: 3.6338. Result ID: 69996cda5698665ec4da9b54. [Reporting paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"8"}. This record does not assert independent reproduction. ### Müller and Montufar (2022) L^pi: 3.6338. Result ID: 69996cda5698665ec4da9b61. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"8"}. This record does not assert independent reproduction. ### Müller and Montufar (2022) L^pi: 3.636. Result ID: 69996cda5698665ec4da9b63. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"10"}. This record does not assert independent reproduction. ### Müller and Montufar (2022) L^pi: 3.6364. Result ID: 69996cda5698665ec4da9b62. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"9"}. This record does not assert independent reproduction. ### Memoryless Policy Iteration L^pi: 3.6378. Result ID: 69996cda5698665ec4da9b4d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"10"}. This record does not assert independent reproduction. ### Policy gradient L^pi: 3.6378. Result ID: 69996cda5698665ec4da9b56. [Reporting paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"10"}. This record does not assert independent reproduction. ### Memoryless Policy Iteration L^pi: 3.6392. Result ID: 69996cda5698665ec4da9b4c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"9"}. This record does not assert independent reproduction. ### Policy gradient L^pi: 3.6392. Result ID: 69996cda5698665ec4da9b55. [Reporting paper]() Recorded evaluation configuration: {"\|O\| and \|A\|":"9"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.