ResearchBenchmarksReinforcement Learning on HalfCheetah tight heterogeneous constraints v5 (test)Reinforcement Learning on HalfCheetah tight heterogeneous constraints v5 (test)Follow4,329ReturnDD-SRad1,370.22,138.352,906.53,674.65May 5, 2026Evaluation resultsMethodMethodLinksReturnCV ScoreUtilizationDD-SRadBackbone=TD3, Seeds=5Backbone=TD3, Seeds=52026.054,32912.266.8D-TanhBackbone=TD3, Seeds=5Backbone=TD3, Seeds=52026.054,0007.967.2BoxPre+Backbone=TD3, Seeds=5Backbone=TD3, Seeds=52026.053,64118.760.7SRad-QPBackbone=TD3, Seeds=5Backbone=TD3, Seeds=52026.053,211458.2SRad-StrictBackbone=TD3, Seeds=5Backbone=TD3, Seeds=52026.051,51812.221.1Post(QP)Backbone=TD3, Seeds=5Backbone=TD3, Seeds=52026.051,48470.269.5