Open Access

ARTICLE

An Adaptive Firefly Algorithm for Dependent Task Scheduling in IoT-Fog Computing

Adil Yousif^*

Department of Computer Science, College of Science and Arts-Sharourah, Najran University, Najran Sharourah, 68378, Saudi Arabia

* Corresponding Author: Adil Yousif. Email:

(This article belongs to the Special Issue: Engineering Applications of Discrete Optimization and Scheduling Algorithms)

Computer Modeling in Engineering & Sciences 2025, 142(3), 2869-2892. https://doi.org/10.32604/cmes.2025.059786

Received 17 October 2024; Accepted 20 January 2025; Issue published 03 March 2025

Abstract

The Internet of Things (IoT) has emerged as an important future technology. IoT-Fog is a new computing paradigm that processes IoT data on servers close to the source of the data. In IoT-Fog computing, resource allocation and independent task scheduling aim to deliver short response time services demanded by the IoT devices and performed by fog servers. The heterogeneity of the IoT-Fog resources and the huge amount of data that needs to be processed by the IoT-Fog tasks make scheduling fog computing tasks a challenging problem. This study proposes an Adaptive Firefly Algorithm (AFA) for dependent task scheduling in IoT-Fog computing. The proposed AFA is a modified version of the standard Firefly Algorithm (FA), considering the execution times of the submitted tasks, the impact of synchronization requirements, and the communication time between dependent tasks. As IoT-Fog computing depends mainly on distributed fog node servers that receive tasks in a dynamic manner, tackling the communications and synchronization issues between dependent tasks is becoming a challenging problem. The proposed AFA aims to address the dynamic nature of IoT-Fog computing environments. The proposed AFA mechanism considers a dynamic light absorption coefficient to control the decrease in attractiveness over iterations. The proposed AFA mechanism performance was benchmarked against the standard Firefly Algorithm (FA), Puma Optimizer (PO), Genetic Algorithm (GA), and Ant Colony Optimization (ACO) through simulations under light, typical, and heavy workload scenarios. In heavy workloads, the proposed AFA mechanism obtained the shortest average execution time, 968.98 ms compared to 970.96, 1352.87, 1247.28, and 1773.62 of FA, PO, GA, and ACO, respectively. The simulation results demonstrate the proposed AFA’s ability to rapidly converge to optimal solutions, emphasizing its adaptability and efficiency in typical and heavy workloads.

---

Keywords

---