Matches in SemOpenAlex for { <https://semopenalex.org/work/W3097232108> ?p ?o ?g. }
- W3097232108 abstract "In neural combinatorial optimization (CO), reinforcement learning (RL) can turn a deep neural net into a fast, powerful heuristic solver of NP-hard problems. This approach has a great potential in practical applications because it allows near-optimal solutions to be found without expert guides armed with substantial domain knowledge. We introduce Policy Optimization with Multiple Optima (POMO), an end-to-end approach for building such a heuristic solver. POMO is applicable to a wide range of CO problems. It is designed to exploit the symmetries in the representation of a CO solution. POMO uses a modified REINFORCE algorithm that forces diverse rollouts towards all optimal solutions. Empirically, the low-variance baseline of POMO makes RL training fast and stable, and it is more resistant to local minima compared to previous approaches. We also introduce a new augmentation-based inference method, which accompanies POMO nicely. We demonstrate the effectiveness of POMO by solving three popular NP-hard problems, namely, traveling salesman (TSP), capacitated vehicle routing (CVRP), and 0-1 knapsack (KP). For all three, our solver based on POMO shows a significant improvement in performance over all recent learned heuristics. In particular, we achieve the optimality gap of 0.14% with TSP100 while reducing inference time by more than an order of magnitude." @default.
- W3097232108 created "2020-11-09" @default.
- W3097232108 creator A5012323997 @default.
- W3097232108 creator A5041200899 @default.
- W3097232108 creator A5061301511 @default.
- W3097232108 creator A5062265168 @default.
- W3097232108 creator A5071320319 @default.
- W3097232108 creator A5081601034 @default.
- W3097232108 date "2020-10-30" @default.
- W3097232108 modified "2023-09-27" @default.
- W3097232108 title "POMO: Policy Optimization with Multiple Optima for Reinforcement Learning" @default.
- W3097232108 cites W1993411524 @default.
- W3097232108 cites W2017708378 @default.
- W3097232108 cites W2097117768 @default.
- W3097232108 cites W2119717200 @default.
- W3097232108 cites W2130942839 @default.
- W3097232108 cites W2145339207 @default.
- W3097232108 cites W2156737235 @default.
- W3097232108 cites W2163605009 @default.
- W3097232108 cites W2183341477 @default.
- W3097232108 cites W2299115575 @default.
- W3097232108 cites W2560592986 @default.
- W3097232108 cites W2607264901 @default.
- W3097232108 cites W2912555327 @default.
- W3097232108 cites W2948433391 @default.
- W3097232108 cites W2962742544 @default.
- W3097232108 cites W2962979969 @default.
- W3097232108 cites W2963341956 @default.
- W3097232108 cites W2963403868 @default.
- W3097232108 cites W2964121744 @default.
- W3097232108 cites W2964308564 @default.
- W3097232108 cites W2966628687 @default.
- W3097232108 cites W2970706905 @default.
- W3097232108 cites W2990626178 @default.
- W3097232108 cites W2996246179 @default.
- W3097232108 cites W3014847873 @default.
- W3097232108 cites W3021435853 @default.
- W3097232108 cites W3047863327 @default.
- W3097232108 cites W3097400948 @default.
- W3097232108 cites W626292722 @default.
- W3097232108 hasPublicationYear "2020" @default.
- W3097232108 type Work @default.
- W3097232108 sameAs 3097232108 @default.
- W3097232108 citedByCount "2" @default.
- W3097232108 countsByYear W30972321082021 @default.
- W3097232108 crossrefType "posted-content" @default.
- W3097232108 hasAuthorship W3097232108A5012323997 @default.
- W3097232108 hasAuthorship W3097232108A5041200899 @default.
- W3097232108 hasAuthorship W3097232108A5061301511 @default.
- W3097232108 hasAuthorship W3097232108A5062265168 @default.
- W3097232108 hasAuthorship W3097232108A5071320319 @default.
- W3097232108 hasAuthorship W3097232108A5081601034 @default.
- W3097232108 hasConcept C113138325 @default.
- W3097232108 hasConcept C11413529 @default.
- W3097232108 hasConcept C126255220 @default.
- W3097232108 hasConcept C127705205 @default.
- W3097232108 hasConcept C141934464 @default.
- W3097232108 hasConcept C154945302 @default.
- W3097232108 hasConcept C173801870 @default.
- W3097232108 hasConcept C175859090 @default.
- W3097232108 hasConcept C2778770139 @default.
- W3097232108 hasConcept C33923547 @default.
- W3097232108 hasConcept C41008148 @default.
- W3097232108 hasConcept C52692508 @default.
- W3097232108 hasConcept C97541855 @default.
- W3097232108 hasConceptScore W3097232108C113138325 @default.
- W3097232108 hasConceptScore W3097232108C11413529 @default.
- W3097232108 hasConceptScore W3097232108C126255220 @default.
- W3097232108 hasConceptScore W3097232108C127705205 @default.
- W3097232108 hasConceptScore W3097232108C141934464 @default.
- W3097232108 hasConceptScore W3097232108C154945302 @default.
- W3097232108 hasConceptScore W3097232108C173801870 @default.
- W3097232108 hasConceptScore W3097232108C175859090 @default.
- W3097232108 hasConceptScore W3097232108C2778770139 @default.
- W3097232108 hasConceptScore W3097232108C33923547 @default.
- W3097232108 hasConceptScore W3097232108C41008148 @default.
- W3097232108 hasConceptScore W3097232108C52692508 @default.
- W3097232108 hasConceptScore W3097232108C97541855 @default.
- W3097232108 hasLocation W30972321081 @default.
- W3097232108 hasOpenAccess W3097232108 @default.
- W3097232108 hasPrimaryLocation W30972321081 @default.
- W3097232108 hasRelatedWork W1490182363 @default.
- W3097232108 hasRelatedWork W1569562017 @default.
- W3097232108 hasRelatedWork W2131431128 @default.
- W3097232108 hasRelatedWork W2406327403 @default.
- W3097232108 hasRelatedWork W2591041104 @default.
- W3097232108 hasRelatedWork W2783833771 @default.
- W3097232108 hasRelatedWork W2791709949 @default.
- W3097232108 hasRelatedWork W2892108814 @default.
- W3097232108 hasRelatedWork W2913564078 @default.
- W3097232108 hasRelatedWork W2948058593 @default.
- W3097232108 hasRelatedWork W2952608458 @default.
- W3097232108 hasRelatedWork W2953681220 @default.
- W3097232108 hasRelatedWork W2957148222 @default.
- W3097232108 hasRelatedWork W2963809395 @default.
- W3097232108 hasRelatedWork W3161231593 @default.
- W3097232108 hasRelatedWork W3187524524 @default.
- W3097232108 hasRelatedWork W3190612470 @default.
- W3097232108 hasRelatedWork W43777471 @default.
- W3097232108 hasRelatedWork W871403656 @default.