# A Change-Detection Based Thompson Sampling Framework for Non-Stationary Bandits

arXiv: https://arxiv.org/abs/2009.02791 · HTML: https://mlsanity.com/paper/2009.02791/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 6 Sep 2020.

## Built on by (most cited first, 10 of 10)

- [Continual Learning as Computationally Constrained Reinforcement Learning](https://mlsanity.com/paper/2307.04345/) (2023)
- [Non-Stationary Bandit Learning via Predictive Sampling](https://mlsanity.com/paper/2205.01970/) (2022)
- [An Information-Theoretic Analysis of Nonstationary Bandit Learning](https://mlsanity.com/paper/2302.04452/) (2023)
- [A Definition of Non-Stationary Bandits](https://mlsanity.com/paper/2302.12202/) (2023)
- [Kolmogorov-Smirnov Test-Based Actively-Adaptive Thompson Sampling for Non-Stationary Bandits](https://mlsanity.com/paper/2105.14586/) (2021)
- [Using Non-Stationary Bandits for Learning in Repeated Cournot Games with Non-Stationary Demand](https://mlsanity.com/paper/2201.00486/) (2022)
- [Fast Change Identification in Multi-Play Bandits and its Applications in Wireless Networks](https://mlsanity.com/paper/2205.10366/) (2022)
- [Non-Stationary Contextual Bandit Learning via Neural Predictive Ensemble Sampling](https://mlsanity.com/paper/2310.07786/) (2023)
- [Safe Sequential Optimization for Switching Environments](https://mlsanity.com/paper/2311.02119/) (2023)
- [Fooling Algorithms in Non-Stationary Bandits using Belief Inertia](https://mlsanity.com/paper/2511.05620/) (2025)

## Abstract (arXiv)

We consider a non-stationary two-armed bandit framework and propose a change-detection based Thompson sampling (TS) algorithm, named TS with change-detection (TS-CD), to keep track of the dynamic environment. The non-stationarity is modeled using a Poisson arrival process, which changes the mean of the rewards on each arrival. The proposed strategy compares the empirical mean of the recent rewards of an arm with the estimate of the mean of the rewards from its history. It detects a change when the empirical mean deviates from the mean estimate by a value larger than a threshold. Then, we characterize the lower bound on the duration of the time-window for which the bandit framework must remain stationary for TS-CD to successfully detect a change when it occurs. Consequently, our results highlight an upper bound on the parameter for the Poisson arrival process, for which the TS-CD achieves asymptotic regret optimality with high probability. Finally, we validate the efficacy of TS-CD by testing it for edge-control of radio access technique (RAT)-selection in a wireless network. Our results show that TS-CD not only outperforms the classical max-power RAT selection strategy but also other actively adaptive and passively adaptive bandit algorithms that are designed for non-stationary environments.