Fog computing, an extension of cloud computing, enhances capabilities by processing data closer to the source, thereby addressing latency and bandwidth issues inherent in traditional cloud models. However, the integration of Artificial Intelligence (AI) into fog computing introduces challenges, particularly in resource management, security, and privacy. This paper systematically reviews AI applications within fog computing environments, adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology to ensure rigorous analysis. The studies were selected based on predefined inclusion criteria, including research published between 2010 and 2024 in peer-reviewed journals and conference papers, with searches conducted in databases like IEEE Xplore, ACM Digital Library, SpringerLink, and Scopus. The review identifies critical issues such as resource constraints, transparency in AI-driven security systems, and the need for adaptable AI models to address evolving security threats. In response, innovative solutions such as lightweight AI models (e.g., Pruned Neural Networks, Quantized Models, Knowledge Distillation), Explainable AI (XAI) (e.g., Model-Agnostic Methods, Feature Importance Analysis, Rule-Based Approaches), and federated learning are proposed. Additionally, a novel taxonomy is introduced, categorizing AI techniques into resource management, security enhancement, and privacy-preserving methods, offering a structured framework for researchers and practitioners. The paper concludes that effective AI integration in fog computing is essential for developing secure, efficient, and adaptable distributed systems, with significant implications for both academia and industry.

Background. Fog computing (FC), extending cloud services to the network edge, has emerged as a key paradigm for low-latency applications like the Internet of Things (IoT). However, efficient resource management, task scheduling, and load balancing pose challenges in fog environments. This review surveys recent research efforts aimed at addressing these challenges and optimizing FC performance. Methodology.We conducted a systematic analysis of relevant research papers on FC published in reputable academic databases. The review focused on studies published between 2019 and 2024 and emphasized artificial intelligence based studies exploring resource management, task scheduling, and load balancing techniques within the FC domain. Results. The review identifies a diverse range of techniques applied to optimize FC performance. These include machine learning (ML) and deep learning (DL) for resource allocation, heuristic algorithms for task scheduling, and nature-inspired meta-heuristics for load balancing. The review evaluates the strengths and limitations of these approaches, highlighting their impact on metrics like latency, energy consumption, and Quality of Service (QoS). Conclusions. This review demonstrates the significant progress made in optimizing FC through innovative techniques. ML and meta-heuristics have emerged as promising approaches for resource management, task scheduling, and load balancing, respectively. However, challenges persist in areas like real-world implementation complexities and ensuring service quality across geographically distributed fog networks. Future research directions are identified, emphasizing the need for further exploration of these challenges and the integration of emerging technologies like deep reinforcement learning for enhanced FC performance.

Cloud computing is considered to be the best solution for addressing the increasing computing requirements of high-performance applications. The efficient performance of the system requires optimal mapping of cloud tasks to resources. However, it is challenging to address computing and storage requirements of high-performance applications while achieving conflicting scheduling objectives like throughput, makespan, resource utilization. This work proposes a metaheuristic approach called task schedule using a multi-objective grey wolf optimizer (TSMGWO) to find near-optimal task scheduling solutions while handling conflicting objectives. The TSMGWO approach has been evaluated using three benchmark datasets, namely, GoCJ, HCSP and Synthetic dataset. The results are compared with heuristic FCFS and MT methods, and metaheuristic methods PSO, GA and WOA. The TSMGWO approach reduces makespan upto 67.52% over FCFS method, 60.93%, over PSO, 38.05% over GA, and 23.22% over WOA methods for 100 tasked cloud workload using GoCJ dataset. It reduces makespan upto 60.95% over FCFS method, 55.79% over MT method, 47.04% over PSO, 33.38% over GA and 19.91% over WOA method using synthetic dataset. Similarly, TSMGWO reduces makespan upto 27.03% over MT method, 18.95% over PSO, 11.90% over GA and 7.5% over WOA method using HSCP workload. The comparative analysis demonstrates that TSMGWO approach outperforms the earlier heuristic and metaheuristic methods using benchmark datasets in the cloud environment.

Congenital heart disease (CHD) represents a multifaceted medical condition that requires early detection and diagnosis for effective management, given its diverse presentations and subtle symptoms that manifest from birth. This research article introduces a groundbreaking healthcare application, the Machine Learning-based Congenital Heart Disease Prediction Method (ML-CHDPM), tailored to address these challenges and expedite the timely identification and classification of CHD in pregnant women. The ML-CHDPM model leverages state-of-the-art machine learning techniques to categorize CHD cases, taking into account pertinent clinical and demographic factors. Trained on a comprehensive dataset, the model captures intricate patterns and relationships, resulting in precise predictions and classifications. The evaluation of the model’s performance encompasses sensitivity, specificity, accuracy, and the area under the receiver operating characteristic curve. Remarkably, the findings underscore the ML-CHDPM’s superiority across six pivotal metrics: accuracy, precision, recall, specificity, false positive rate (FPR), and false negative rate (FNR). The method achieves an average accuracy rate of 94.28%, precision of 87.54%, recall rate of 96.25%, specificity rate of 91.74%, FPR of 8.26%, and FNR of 3.75%. These outcomes distinctly demonstrate the ML-CHDPM’s effectiveness in reliably predicting and classifying CHD cases. This research marks a significant stride toward early detection and diagnosis, harnessing advanced machine learning techniques within the realm of ECG signal processing, specifically tailored to pregnant women.

Fog computing has emerged as a prospective paradigm to address the computational requirements of IoT applications, extending the capabilities of cloud computing to the network edge. Task scheduling is pivotal in enhancing energy efficiency, optimizing resource utilization and ensuring the timely execution of tasks within fog computing environments. This article presents a comprehensive review of the advancements in task scheduling methodologies for fog computing systems, covering priority-based, greedy heuristics, metaheuristics, learning-based, hybrid heuristics, and nature-inspired heuristic approaches. Through a systematic analysis of relevant literature, we highlight the strengths and limitations of each approach and identify key challenges facing fog computing task scheduling, including dynamic environments, heterogeneity, scalability, resource constraints, security concerns, and algorithm transparency. Furthermore, we propose future research directions to address these challenges, including the integration of machine learning techniques for real-time adaptation, leveraging federated learning for collaborative scheduling, developing resource-aware and energy-efficient algorithms, incorporating security-aware techniques, and advancing explainable AI methodologies. By addressing these challenges and pursuing these research directions, we aim to facilitate the development of more robust, adaptable, and efficient task-scheduling solutions for fog computing environments, ultimately fostering trust, security, and sustainability in fog computing systems and facilitating their widespread adoption across diverse applications and domains.

Optimal allocation of virtual machines in a cloud computing environment for user-submitted tasks is a challenging task. Finding an optimal task scheduling solution is considered as NP-hard problem specifically for large task sizes in the cloud environment. The best solution involves scheduling the tasks to virtual machines data centre while minimizing the essential, influential and cost effective parameters such as energy usage, makespan and cost. In this direction, this work presents a metaheuristic framework called MDVMA for dynamic virtual machine allocation with optimized task scheduling in a cloud computing environment. The MDVMA focuses on developing a multi-objective scheduling method using non dominated sorting genetic algorithm (NSGA)-II algorithm-based metaheuristic algorithm for optimizing task scheduling with the aim of minimizing energy usage, makespan and cost simultaneously to provide trade-off to the cloud service providers as per their requirements. To evaluate the performance of the MDVMA approach, we compared the performances of two different scenarios of benchmark real-world workload data sets using the existing approaches, namely, Artificial Bee Colony (ABC) algorithm, Whale Optimization Algorithm (WOA) and Particle Swarm Optimization (PSO) algorithm. Simulation results demonstrate that optimizing task scheduling leads to better overall results in terms of minimizing energy usage, makespan and cost of the cloud data center. Finally, the paper concludes metaheuristic algorithms as a promising method for task scheduling in a cloud computing environment.