English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Reinforcement Learning (RL) for constrained MDPs (CMDPs) is an increasinglyimportant problem for various applications. Often, the average criterion ismore suitable than the discounted criterion. Yet, RL for average-CMDPs (ACMDPs)remains a challenging problem. Algorithms designed for discounted constrainedRL problems often do not perform well for the average CMDP setting. In thispaper, we introduce a new policy optimization with function approximationalgorithm for constrained MDPs with the average criterion. TheAverage-Constrained Policy Optimization (ACPO) algorithm is inspired by trustregion-based policy optimization algorithms. We develop basic sensitivitytheory for average CMDPs, and then use the corresponding bounds in the designof the algorithm. We provide theoretical guarantees on its performance, andthrough extensive experimental work in various challenging OpenAI Gymenvironments, show its superior empirical performance when compared to otherstate-of-the-art algorithms adapted for the ACMDPs.

ACPO: A Policy Optimization Algorithm for Average MDPs with Constraints