Unverified50% confidenceFactExact time
If the shaping reward takes the form F(s, s') = γΦ(s') − Φ(s), then the optimal policy obtained under this shaping is exactly identical to the optimal policy of the original sparse-reward problem.
1
Sources
50%
Confidence
Long-term
Relevance
10/14/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedPotential-based reward shaping guarantees that the optimal policy remains unchanged.73% similarUnverified吊挂控制的典型奖励函数形如 r = -α|θ_yaw| - β|ω_yaw| - γ||u||² - λ||Δu||²，分别惩罚偏航偏差、旋转角速度、控制输入大小和控制输入变化率63% similarUnverifiedReward shaping refers to adding extra, denser guiding reward signals on top of the original sparse reward to help agents learn faster and more correctly.61% similarUnverifiedActor-Critic架构中用优势函数A(s,a)=Q(s,a)-V(s)替代原始回报G_t，可以在不改变梯度期望的前提下大幅缩小方差60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1024907API
curl https://kongchang.com/api/v1/knowledge/claims/1024907MCP
get_claim(id=1024907)