Online data assimilation in distributionally robust optimization

TL;DR

This paper introduces an online data assimilation algorithm for distributionally robust optimization that guarantees high-probability out-of-sample performance and improves decision quality as more streaming data is collected.

Contribution

It proposes a novel online algorithm that adaptively incorporates streaming data into distributionally robust optimization, ensuring convergence and performance guarantees.

Findings

Algorithm guarantees high-probability out-of-sample performance.

Decentralized convergence of the data assimilation process.

Provides criteria for algorithm termination based on data collection.

Abstract

This paper considers a class of real-time decision making problems to minimize the expected value of a function that depends on a random variable $\xi$ under an unknown distribution $\mathbb{P}$. In this process, samples of $\xi$ are collected sequentially in real time, and the decisions are made, using the real-time data, to guarantee out-of-sample performance. We approach this problem in a distributionally robust optimization framework and propose a novel Online Data Assimilation Algorithm for this purpose. This algorithm guarantees the out-of-sample performance in high probability, and gradually improves the quality of the data-driven decisions by incorporating the streaming data. We show that the Online Data Assimilation Algorithm guarantees convergence under the streaming data, and a criteria for termination of the algorithm after certain number of data has been collected.