Asymptotically Optimal Anomaly Detection via Sequential Testing

TL;DR

This paper develops asymptotically optimal sequential testing algorithms for detecting anomalies among multiple processes with limited observations, minimizing total expected costs while satisfying error constraints.

Contribution

It introduces low-complexity algorithms that achieve asymptotic optimality in anomaly detection with sequential testing under resource and reliability constraints.

Findings

Algorithms are asymptotically optimal as error constraints approach zero.

Simulation results show strong finite-sample performance.

The method effectively balances observation costs and detection accuracy.

Abstract

Sequential detection of independent anomalous processes among K processes is considered. At each time, only M processes can be observed, and the observations from each chosen process follow two different distributions, depending on whether the process is normal or abnormal. Each anomalous process incurs a cost per unit time until its anomaly is identified and fixed. Switching across processes and state declarations are allowed at all times, while decisions are based on all past observations and actions. The objective is a sequential search strategy that minimizes the total expected cost incurred by all the processes during the detection process under reliability constraints. Low-complexity algorithms are established to achieve asymptotically optimal performance as the error constraints approach zero. Simulation results demonstrate strong performance in the finite regime.