Markov State Assumptions in Optimal Execution Models

Summary

The document raises a modeling question about reinforcement learning for optimal execution. It asks whether the Markov property commonly assumed in this literature is justified, and whether empirical work compares non-Markovian models with fewer state variables against Markov models with richer states.

It does not provide a proposed method, empirical results, or a resolution. The cited papers are examples of the research context, while the main contribution is identifying a tradeoff for investigation: state richness and Markov assumptions may affect model performance. The question alone cannot establish which formulation is better, and no specific execution setting or evaluation criteria are given.

Key ideas

• The document questions whether Markov state assumptions are appropriate for reinforcement learning in optimal execution.
• It asks whether compact non-Markovian models have been empirically compared with richer Markov models.
• No comparison results or recommended state design are provided.

Tags

# Markov Property in Optimal Execution?


# Markov Property in Optimal Execution?












After reading papers on reinforcement learning with respect to the problem of optimal execution (Nevmyvaka et al (2006), Ning et al (2018), etc), I was wondering if the Markov property assumed in all these papers is rational? Has there been at least an empirical comparison/paper detailing tradeoffs & performance of non-Markovian models with less state variables vs a Markov model with more state variables someone could direct me towards?

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.