{"id":"W4379523790","doi":"10.2118/213104-ms","title":"Reinforcement Learning for Multi-Well SAGD Optimization: A Policy Gradient Approach","year":2023,"lang":"en","type":"article","venue":"","topic":"Reservoir Engineering and Simulation Methods","field":"Engineering","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Reinforcement learning; Markov decision process; Computer science; Mathematical optimization; Process (computing); Convergence (economics); Artificial neural network; Markov process; Action (physics); Monte Carlo method; Bellman equation; Artificial intelligence; Mathematics","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"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":[0.002119993,0.001014204,0.002063082,0.0007699548,0.0004063687,0.001039263,0.001273843,0.001668417,0.003362291],"category_scores_gemma":[0.003842968,0.0006672156,0.0006697857,0.0005060819,0.001408464,0.0007831794,0.001551226,0.001564298,0.0003362284],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00142123,"about_ca_system_score_gemma":0.001435259,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007476482,"about_ca_topic_score_gemma":0.003553173,"domain_scores_codex":[0.9995421,0.000220076,0.00001890627,0.00006260488,0.00009562058,0.000060754],"domain_scores_gemma":[0.99822,0.001202422,0.0001263368,0.00004776188,0.0003097809,0.00009361217],"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":[0.00002060029,0.00002015854,0.0002157378,0.00003081985,0.00001657289,0.0000349458,0.00001512651,0.9882683,0.0001615865,0.005654471,0.0002614449,0.005300246],"study_design_scores_gemma":[0.000003619002,0.000006851552,0.00001465422,0.000003172393,0.000001486599,0.000001597908,0.000001825568,0.9985532,0.0000222669,0.001294468,0.00009556562,0.00000133351],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01944418,0.0007675118,0.9730684,0.000844463,0.00007586752,0.00008208532,0.00004505405,0.0003056916,0.005366925],"genre_scores_gemma":[0.8787522,0.0004714724,0.1145982,0.0003617608,0.0001074101,0.0003319217,0.0001053346,0.0001183104,0.005153473],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007476482,"threshold_uncertainty_score":0.01486593,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05155217862445684,"score_gpt":0.3101497248625225,"score_spread":0.2585975462380657,"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."}}