Periodic-GP: Learning Periodic World with Gaussian Process Bandits
This addresses optimization problems with seasonality, such as ride-sharing demand or traffic patterns, but is incremental as it builds on existing GP bandit frameworks.
The paper tackles sequential decision optimization in periodic environments by proposing Periodic-GP, a Gaussian process bandit method with a temporal periodic kernel, which achieves a new theoretical regret bound and significantly outperforms existing methods in synthetic and real-world traffic pollution data experiments.
We consider the sequential decision optimization on the periodic environment, that occurs in a wide variety of real-world applications when the data involves seasonality, such as the daily demand of drivers in ride-sharing and dynamic traffic patterns in transportation. In this work, we focus on learning the stochastic periodic world by leveraging this seasonal law. To deal with the general action space, we use the bandit based on Gaussian process (GP) as the base model due to its flexibility and generality, and propose the Periodic-GP method with a temporal periodic kernel based on the upper confidence bound. Theoretically, we provide a new regret bound of the proposed method, by explicitly characterizing the periodic kernel in the periodic stationary model. Empirically, the proposed algorithm significantly outperforms the existing methods in both synthetic data experiments and a real data application on Madrid traffic pollution.