ResearchBenchmarksConstrained Offline Reinforcement Learning on DSRL PointCircle1Constrained Offline Reinforcement Learning on DSRL PointCircle1Follow0.86Normalized ReturnCOptiDICE0.3920.51350.6350.7565Dec 9, 2025Evaluation resultsMethodMethodLinksNormalized ReturnCostCOptiDICE2025.120.865.51BC-All2025.120.793.98CDT2025.120.590.69MPDiffusernum_samples=16num_samples=162025.120.580.94BCQ-Lag2025.120.542.38CPQ2025.120.430.75BC-Safe2025.120.410.16