Tackling Decision Processes with Non-Cumulative Objectives using Reinforcement Learning
By Maximilian Nägele, Jan Olle, Thomas Fösel, Remmy Zen, Florian Marquardt
Rating
1379
Battle Count: 70
Relevance
8/10
Directly applicable to portfolio optimization with Sharpe ratio objective, potentially improving both performance and training time
Implementation Complexity
5/10
Requires minimal changes to existing reinforcement learning implementations, but understanding the theoretical framework may be challenging
Reproducibility
4/5
Code and data are open-sourced, detailed experimental setup provided
About this paper
Methodology: Non-Cumulative Markov Decision Process Mapping. Problem types: Reinforcement Learning, Portfolio Optimization, Discrete Optimization, Quantum Error Correction.
The interactive Everscope explorer (charts, battles, favorites) loads below.