In order to solve the problems of effective resource allocation in low-power wide-area networks, this thesis investigates the scheduling of end devices in Internet of Things applications using LoRaWAN technology. The main goal of this research is to use RL to improve QoS measures including energy efficiency, throughput, latency, and dependability. This was accomplished by using a simulation-based approach that evaluated the effectiveness of the RL-based scheduling algorithm using NS3 simulations. The main findings show that, in comparison to current scheduling practices, the RL agent greatly improves data transmission reliability and improves network throughput. Furthermore, the suggested approach efficiently lowers average system latency and overall energy usage, improving network resource utilization. These findings imply that using reinforcement learning (RL) for job scheduling in LoRaWAN networks can offer a reliable and expandable solution to present problems, resulting in more intelligent and environmentally friendly IoT systems. In the end, this study finds that using RL-based techniques can help improve resource management in contexts that are dynamic and resource-constrained.