A CVD Critical Level-aware Scheduling Model Based on Reinforcement Learning for ECG Service Request

Abstract

In the cardiovascular disease (CVD) diagnosis scenario, the number of electrocardiogram (ECG) service request data is large and the severity of CVD is different. Efficient task scheduling is the key to large cluster computer-aided CVD diagnosis. Therefore, in task scheduling, the workload changes and the critical condition of CVD must be paid attention to. We propose a CVD critical level-aware scheduling model based on reinforcement learning (CLS-RL) to optimize ECG service request scheduling. To solve the problem that there is no publicly available ECG service request data, this paper proposes a method of composing it. Then, we utilize RL with Actor-Critic to improve the efficiency of scheduling. Finally, we define the new objective functions for ECG service request scheduling. The experimental results show that the proposed CLS-RL is the best in comprehensive performance.