Enhancing Video Streaming Quality With Reinforcement Learning‐Based Fuzzy Crossover Mountain‐Gazelle Algorithm

The process of streaming efficiency through adaptive bitrate control can reduce unnecessary bandwidth usage while maintaining perceived video quality and minimize the occurrence of buffering over the Internet. There are a variety of solutions available today for improving the overall quality of service associated with video streaming; however, there are still many obstacles to achieving these objectives, such as excessive processing power requirements or costs that are too high. Some currently available solutions do not produce the desired results because there is an insufficient amount of important detail available; therefore, there is a lack of innovative solutions designed for increasing end‐user satisfaction. Therefore, this work develops a novel approach called the Reinforcement Learning‐based Fuzzy Crossover Mountain‐Gazelle algorithm in improving the quality of service of video streaming to optimize client‐side and server‐side cooperation. The reinforcement learning method is applied in this model's design to improve the user experience of streaming video over adjustable network conditions and facilitate adaptive decision‐making. Fuzzy logic handles uncertainty in network conditions and maps imprecise quality of service indicators into quality of experience‐oriented decision rules. Additionally, other hyperparameter tuning to improve the overall performance of the model will be achieved using a crossover‐based mountain‐gazelle algorithm. Extensive testing of the developed solution is performed using various testing parameters to evaluate the developed solution relative to these testing parameters. Moreover, through experimental validation, the proposed method acquires the best performance with a bandwidth of 15 Mbps. This specifies the superior performance for higher user satisfaction compared to the other conventional methods.