Abstract:Goal-conditioned reinforcement learning (RL) concerns the problem of training an agent to maximize the probability of reaching target goal states. This paper presents an analysis of the goal-conditioned setting based on optimal control. In particular, we derive an optimality gap between more classical, often quadratic, objectives and the goal-conditioned reward, elucidating the success of goal-conditioned RL and why classical ``dense'' rewards can falter. We then consider the partially observed Markov decision setting and connect state estimation to our probabilistic reward, making the goal-conditioned reward well suited to dual control problems. The advantages of goal-conditioned policies are validated on nonlinear and uncertain environments using both RL and predictive control techniques.
| Comments: | IFAC world congress postprint |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2512.06471 [cs.LG] |
| (or arXiv:2512.06471v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2512.06471 arXiv-issued DOI via DataCite |
Submission history
From: Nathan P. Lawrence [view email]
[v1]
Sat, 6 Dec 2025 15:28:35 UTC (245 KB)
[v2]
Thu, 14 May 2026 17:36:34 UTC (243 KB)