ResearchBenchmarksOffline Control on Heterogeneous Pendulum 300,000 transition steps (Mid-Data)Follow-1.25Cumulative RewardAOC-718.0596-531.9648-345.87-159.7752Oct 11, 2023Evaluation ResultsMethodMethodLinksCumulative RewardAOC2023.10-1.25MPC2023.10-1.34MFRL2023.10-11.95BC2023.10-225.32Data-Avg-Return2023.10-245.54KNN2023.10-670.511NN2023.10-690.49