ResearchBenchmarksOffline Control on Heterogeneous Pendulum Rich-Data 600,000 transition stepsFollow-0.6Cumulative RewardAOC-671.5248-497.3424-323.16-148.9776Oct 11, 2023Evaluation ResultsMethodMethodLinksCumulative RewardAOC2023.10-0.6MPC2023.10-1.41MFRL2023.10-15.27BC2023.10-126.74Data-Avg-Return2023.10-208.811NN2023.10-512.71KNN2023.10-645.72