Optimizing Mobile Edge Computing Efficiency through Multidimensional Generative Adversarial Network

Big data Processing and the IoT have recently gained popularity as study topics in MEC and thorough research is sought for well-informed decisions. Mobile edge equipment, such as base stations, routers, and edge servers, usually have limited processing capabilities compared to traditional data centers. According to this perspective, this research constructs MEC using Intelligent Outlier Detection with a Multidimensional Generative Adversarial Network (IOD-MDGAN). The suggested structure uses servers at the edge, mobile phones, and cloud resources to achieve minimal latency, lower bandwidth usage, and more scalability. To eliminate outliers from the data, the suggested model uses an adaptive synthetic sampling-based outlier identification technique. Complete start and completion timings for base station connectivity are included in the statistics to gather the information for every mobile user. Data pre-processing for the Adaptive Median Filter (AMF) filter eliminates the noisy data from raw data samples. Feature selection was used for Particle Swarm Optimization (PSO) to choose a suitable group of characteristics. To differentiate between several class labels, an extended immediate memory-based classification model is needed. The innovative nature of the study is demonstrated by the designs of the PSO algorithm for feature selection with the Synthetic Minority Over-sampling Technique (SMOTE) approach for massive data. The results show that the IOD-MDGAN technique has better latency (78 ms), higher throughput (750 Mbps), a faster detection rate (96 ms), a higher data delivery ratio (900 ms), and a higher cost-effectiveness (79). Adapted big data processing architectures designed for MEC provide significant potential for reshaping data use inside mobile applications.