Recent years have witnessed a significant advancement of networking technologies as well as the proliferation of mobile devices. Due to the... Show moreRecent years have witnessed a significant advancement of networking technologies as well as the proliferation of mobile devices. Due to the convergence of pervasive connectivity and ubiquitous computing, Internet of Things (IoT) systems are becoming increasingly information-centric. For those IoT devices, wireless communication is the dominant way to exchange information. The development of IoT has spawned a plethora of real-time applications, boosting the demand for timely information updates. Age of Information (AoI) has recently been introduced to quantify the freshness of the knowledge the controller has about the remote information sources. Due to its sheer novelty in capturing the timeliness requirements of various applications, AoI has sparked tremendous interest and been studied in many communication systems. This thesis aims at an exploratory study on how to characterize the essence of wireless scheduling for effective information freshness from the decision-making perspectives through two representative application scenarios, information retrieval and information integration. For the former, request-aware proactive scheduling policies in both static and dynamic request patterns are developed, which target at minimizing time-average effective AoI (EAoI). For the latter, an experience-driven scheduling framework based on deep reinforcement learning techniques is investigated to minimize the time-average AoI in the presence of correlated information sources. Future research directions are also discussed to present possible extensions of this thesis work to a broader range of network scenarios. Show less
