{"id":"W4415300083","doi":"10.1080/21680566.2025.2561863","title":"Learning heuristics for transit network design and improvement with deep reinforcement learning","year":2025,"lang":"en","type":"article","venue":"Transportmetrica B Transport Dynamics","topic":"Transportation Planning and Optimization","field":"Social Sciences","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Reinforcement learning; Heuristics; Learning network; Network planning and design; Public transport; Transit (satellite)","routes":{"ca_aff":true,"ca_fund":true,"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":[0.001579623,0.001548788,0.001109703,0.001030018,0.0004826866,0.0009547294,0.001392306,0.001187849,0.003086292],"category_scores_gemma":[0.006531401,0.0007821341,0.0006401037,0.0007478687,0.001386428,0.001447713,0.001031423,0.001641344,0.0003239265],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00223193,"about_ca_system_score_gemma":0.002200974,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01053912,"about_ca_topic_score_gemma":0.0165633,"domain_scores_codex":[0.9994711,0.0001834735,0.00002844973,0.0001296996,0.00009053101,0.00009667805],"domain_scores_gemma":[0.9969929,0.002155432,0.0002741663,0.0001573941,0.0002870229,0.0001329485],"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":[0.00002953853,0.00003181807,0.0004355617,0.00003488212,0.00001740755,0.00002335184,0.00002400998,0.9801481,0.0002832414,0.003818531,0.0005231268,0.01463055],"study_design_scores_gemma":[0.00001009672,0.00001469908,0.00003536413,0.00000599066,0.00000463562,0.000003494278,0.000006162431,0.9966398,0.0001302967,0.002975863,0.0001715993,0.000001854767],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06568325,0.0008562962,0.9265029,0.0005558897,0.00007155143,0.0001264138,0.0001263019,0.0009004889,0.005176863],"genre_scores_gemma":[0.8322554,0.0002947311,0.1647725,0.0002557371,0.00003351886,0.000162493,0.0002050471,0.0001387781,0.001881725],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01053912,"threshold_uncertainty_score":0.02095556,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.009609201457943787,"score_gpt":0.246766158077909,"score_spread":0.2371569566199652,"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."}}