{"id":"W4401305302","doi":"10.1016/j.knosys.2024.112312","title":"A safe reinforcement learning algorithm for supervisory control of power plants","year":2024,"lang":"en","type":"article","venue":"Knowledge-Based Systems","topic":"Advanced Control Systems Optimization","field":"Engineering","cited_by":11,"is_retracted":false,"has_abstract":false,"ca_institutions":"Microsoft (Canada)","funders":"Argonne National Laboratory; Laboratory Directed Research and Development","keywords":"Reinforcement learning; Supervisory control; Computer science; Lagrange multiplier; Representation (politics); Control (management); State (computer science); Relaxation (psychology); Control theory (sociology); Optimal control; Mathematical optimization; Power (physics); Lagrangian relaxation; Control engineering; Artificial intelligence; Algorithm; Engineering; Mathematics","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":[0.001100817,0.000611999,0.0007988453,0.0003700361,0.0005043698,0.0005639715,0.001053625,0.0009721025,0.002114206],"category_scores_gemma":[0.002748592,0.0003651708,0.0003934304,0.0002264929,0.0009350085,0.0005953889,0.001291463,0.001386873,0.0004158782],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005889702,"about_ca_system_score_gemma":0.001623666,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004525937,"about_ca_topic_score_gemma":0.002980958,"domain_scores_codex":[0.9996611,0.00009205585,0.00001641567,0.00005145539,0.0001253228,0.0000536013],"domain_scores_gemma":[0.9990677,0.0005366631,0.00007036454,0.00006063003,0.0001947119,0.00006993311],"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":[0.0001511381,0.00004726586,0.0002548612,0.00004233185,0.00002632905,0.00006001439,0.00006546813,0.897902,0.002464918,0.01476145,0.0008448129,0.08337943],"study_design_scores_gemma":[0.00001844665,0.00003159116,0.00002389901,0.000002896609,0.000003023797,0.000006891933,0.000002072219,0.9959862,0.0003930872,0.003315737,0.0002136368,0.000002453112],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.008076677,0.00005689582,0.9906433,0.00006812513,0.00002608746,0.00002196726,0.00001226576,0.0003550039,0.0007396126],"genre_scores_gemma":[0.7247416,0.0001133644,0.2713065,0.0001385729,0.00004144361,0.0001930006,0.0001004254,0.0001351897,0.003229791],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004525937,"threshold_uncertainty_score":0.008999169,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01158204449223214,"score_gpt":0.2260626628308361,"score_spread":0.214480618338604,"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."}}