HLF-D*: An Approximate Greedy Policy for Age-based Information Freshness of Real-time Updates

摘要

In this work, we have considered an industrial cyber-physical system (ICPS) consisting of a symmetric wireless sensor-actuator network (WSAN), controlled by a centralized controller. Real-time sensor updates with variable deadlines are sent to the controller for remote monitoring and closed-loop control of the network. These samples get aperiodically replaced by fresh new samples having variable deadlines, once they fail to get service within their stipulated deadlines. This paper aims to accomplish both the objectives of maximum information freshness and minimum deadline misses of sample packets. For this purpose, a new low complexity greedy scheduling algorithm called Deadline-aware Highest Latency First Advanced (HLF-D*) is proposed in this paper. Through competitive analysis, it is found that the HLF-D* policy has an expected approximation ratio strictly greater than 0.6875. Moreover, the performance of HLF-D* has been compared with other popular deadline-aware scheduling policies through extensive simulations.