Abstract :

Edge–fog Computing Environments Are Increasingly Vital For Handling Large-scale, Real-time Workloads Generated By Digital Services, Internet Of Things (IoT), And Distributed Applications. However, Efficient Task Scheduling Remains A Major Challenge Due To Dynamic Resource Availability, Heterogeneous Architectures, And Diverse Workload Patterns. Traditional Approaches, Such As Static Heuristics And Rulebased Schedulers, As Well As Conventional Machine Learning (ML) Models, Often Fail To Adapt To Changing System Conditions. These Methods Frequently Misclassify Task Requirements, Lack The Ability To Predict Multiple Scheduling Parameters Simultaneously, And Struggle To Scale Effectively, Leading To Reduced System Throughput And Inefficient Resource Utilization. To Address These Limitations, This Study Proposes An Adaptive, Data-driven Scheduling Framework That Combines Preprocessing, Exploratory Data Analysis (EDA), And Multi-model Prediction. The Proposed Convolutional Recurrent Network With Greedy Rule Interpretable Machine (CRN-GRIM) Captures Temporal Workload Patterns While Maintaining Interpretability Through Rule-based Reasoning. Additional Models, Including Passive Aggressive Classifier (PAC), Gaussian Naive Bayes (GNB), And K-Nearest Neighbors Classifier (KNNC), Are Incorporated For Comparative Evaluation And Robustness. The Framework Predicts Multiple Scheduling Targets Are Job Priority, Scheduler Type, And Resource Allocation Within A Unified Multi-output Model, Enhancing Decision Accuracy. A Lightweight Storage System And Web-based Interface Enable Real-time Prediction, Visualization, Monitoring, And Retraining. This Approach Delivers A Scalable, Intelligent, And Efficient Scheduling Solution For Modern Edge–fog Computing Environments.

Published:

09-4-2026

Issue:

Vol. 26 No. 4 (2026)

Page Nos:

2096-2106

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite