{"id":"W3033044509","doi":"10.1609/aaai.v35i5.16484","title":"Combining Reinforcement Learning and Constraint Programming for Combinatorial Optimization","year":2021,"lang":"en","type":"article","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Constraint Satisfaction and Optimization","field":"Computer Science","cited_by":125,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto; Polytechnique Montréal","funders":"","keywords":"Combinatorial optimization; Mathematical optimization; Reinforcement learning; Constraint programming; Knapsack problem; Travelling salesman problem; Computer science; Optimization problem; Solver; Heuristics; Context (archaeology); Quadratic assignment problem; Integer programming; Lin–Kernighan heuristic; Mathematics; Stochastic programming; 2-opt; Artificial intelligence","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002061545,0.001724118,0.001691387,0.0008818653,0.0003844722,0.002001118,0.001878309,0.001626218,0.004227802],"category_scores_gemma":[0.005652525,0.0006790503,0.001091274,0.001491799,0.001451626,0.001625103,0.001766273,0.003384291,0.0006976727],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001351416,"about_ca_system_score_gemma":0.002339954,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006496829,"about_ca_topic_score_gemma":0.007379756,"domain_scores_codex":[0.9987225,0.0005498214,0.00006994542,0.0001990269,0.0003365048,0.0001222924],"domain_scores_gemma":[0.9965966,0.002676298,0.0001742714,0.0001869467,0.0002678764,0.00009799347],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002530124,0.00005626595,0.0003299361,0.0001928627,0.00005277844,0.00006433334,0.0000278423,0.89977,0.0004855022,0.03907723,0.001839957,0.05807795],"study_design_scores_gemma":[0.000006893112,0.0000119695,0.00002665604,0.00001358741,0.000005568776,0.00001061483,0.000003971696,0.9835111,0.0001682861,0.01507982,0.001157333,0.000004186841],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002386039,0.0006039744,0.9926373,0.0003579771,0.00004638525,0.00004632822,0.00003771724,0.0002624873,0.003621799],"genre_scores_gemma":[0.31498,0.002035091,0.6761071,0.0006951977,0.0003196519,0.0005744931,0.0003279856,0.0003352775,0.004625275],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.006496829,"threshold_uncertainty_score":0.01414335,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04400025295225241,"score_gpt":0.2862828622392445,"score_spread":0.2422826092869921,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}