Optimizing Task Scheduling in Fog Computing with Deadline Awareness
Mohammad Sadegh Sirjani, Somayeh Sobati-Moghadam · 2025
Read Optimizing Task Scheduling in Fog Computing with Deadline Awareness by Mohammad Sadegh Sirjani, Somayeh Sobati-Moghadam (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-07378.txt.
Introduction
Mohammad Sadegh Sirjani and 4 co-authors tackle task scheduling in fog computing in a 2025 paper. Fog computing places small servers near Internet of Things devices so time-sensitive jobs get quick responses. The challenge is assigning each task to a node without wasting energy or missing deadlines, the latest times by which jobs must finish.
Nodes are split into low-traffic and high-traffic groups. Short-deadline tasks go to quiet nodes under an improved golden eagle optimization method, a search inspired by hunting eagles and tuned with genetic operators. Long-deadline tasks go to busy nodes under reinforcement learning, where an agent learns from rewards. The combined method, RIGEO, cut energy use by up to 29 percent and response time by up to 86 percent in tests against leading methods, with fewer missed deadlines.
Catalog record
| CREATOR | Mohammad Sadegh Sirjani, Somayeh Sobati-Moghadam |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 30 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-07378
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Optimizing Task Scheduling in Fog Computing with Deadline Awareness Mohammad Sadegh Sirjani, Somayeh Sobati-Moghadam Abstract The rise of Internet of Things (IoT) devices has led to the development of numerous applications that require quick responses and low latency. Fog computing has emerged as a solution for processing these IoT applications, but it faces challenges such as resource allocation and job scheduling. Therefore, it is crucial to determine how to assign and schedule tasks on Fog nodes. A well-designed job scheduling algorithm can help decrease energy usage and improve response times for application requests. This work aims to schedule tasks in IoT while minimizing the total energy consumption of nodes and enhancing the Quality of Service (QoS) requirements of IoT tasks, taking into account task deadlines. Initially, this paper classifies the Fog nodes into two categories based on their traffic level: low and high. It schedules low-deadline tasks on low-traffic-level nodes using an Improved Golden Eagle Optimization (IGEO) algorithm, an enhancement of the Golden Eagle Optimization Algorithm that utilizes genetic operators for discretization. High-deadline tasks are processed on high-traffic nodes using reinforcement learning (RL). This combined approach is called the Reinforcement Improved Golden Eagle Optimization (RIGEO) algorithm. Experimental results demonstrate that the proposed algorithms optimize system response time, total deadline violation time, and resource and system energy consumption compared to other state-of-the-art algorithms. I Introduction In recent years, the Internet of Things (IoT) has become a prominent technology in the internet and networking industry [ 1 ] . It consists of a system of linked physical devices, like household items and cars, that have sensors, software, and online connectivity [ 2 ] . Although IoT has been expanding quickly, it is encountering obstacles such as traffic congestion, delays, and excessive energy usage. In the realm of IoT, a deadline refers to a specific time by which a task or operation must be completed. Adhering to deadlines is crucial in real-time applications where prompt responses are vital, such as e-health in the healthcare sector, smart grid in energy management, livestock monitoring in agriculture, and smart traffic in transportation management. These time-sensitive IoT applications typically have stringent quality of service (QoS) requirements and necessitate processing capabilities that surpass the limitations of IoT devices, which often have significant resource constraints [ 3 ] . Effectively managing resources in IoT and minimizing energy usage while meeting task deadlines can be a challenging task [ 4 ] . Fog computing has emerged as a solution to address challenges in IoT networks by providing computational capabilities at the edge, reducing delays and energy consumption [ 5 , 6 ] . However, using fog presents its own set of challenges, as resources are diverse and limited, necessitating efficient resource scheduling and allocation strategies. Developing practical approaches for allocating resources in fog nodes is essential for optimal performance [ 7 ] . However, despite these advancements, many existing solutions often rely on a single optimization strategy or fail to fully account for the dynamic and heterogeneous nature of real-world IoT-Fog environments, particularly in terms of varying network traffic levels and diverse task deadline requirements [ 8 , 9 , 10 , 11 ] . To address these limitations, this work proposes a novel adaptive task scheduling framework, the Reinforcement Improved Golden Eagle Optimization (RIGEO) algorithm. A key innovation of RIGEO lies in its dynamic classification of Fog Nodes (FNs) into low-traffic and high-traffic categories based on real-time network conditions. Low-traffic nodes are characterized by average network traffic falling below a predefined threshold, while high-traffic nodes are those that exceed this average. This allows for a tailored scheduling approach, leveraging the strengths of different optimization paradigms for distinct operational scenarios. Specifically, for low-deadline tasks on low-traffic nodes, RIGEO employs an enhanced meta-heuristic algorithm, the Improved Golden Eagle Optimization (IGEO), which is designed for rapid and efficient resource allocation to meet stringent time constraints. Conversely, for high-deadline tasks destined for high-traffic nodes, Reinforcement Learning (RL) is utilized, offering robust decision-making capabilities to navigate complex and congested environments while still adhering to deadlines. The proposed RIGEO algorithm operates within the following three-layered architecture: cloud at the top, fog in the middle, and IoT devices at the bottom. IoT devices enable a wide range of devices, including wearable technology, RFID tags, sensors in automobiles and security systems, smart home appliances, thermostats, meters, and industrial equipment. The fog environment consists of a fog controller (FC), also known as a broker, and a network for managing resources and scheduling tasks. The cloud environment includes virtual machines with significant computing and storage capabilities. This architecture is illustrated in Figure 1 . Within this architectural context, since task scheduling has been introduced as an NP-hard problem, conventional methods are unable to solve it [ 12 , 13 ] . Therefore, it is suggested to utilize machine learning and heuristic methods to solve this problem [ 14 ] . Fig. 1: The architecture of the IoT-Fog-Cloud network. There are multiple challenges in the field of task scheduling, including cost optimization, energy efficiency, meeting deadline constraints, reducing latency, ensuring reliability, and achieving fault tolerance. Many studies have been conducted in each of these areas, and some of them will be mentioned in this section. I-A Energy Efficiency Authors in [ 15 ] proposed a new task scheduling algorithm for fog computing environments, called MEETS. The algorithm employs a two-step approach: first, it determines the optimal task allocation to fog nodes, and then it schedules the tasks on each node accordingly. The results show that MEETS can significantly reduce energy consumption while still meeting all task deadlines. However, they considered homogeneous fog networks and assumed that the task arrival rate is constant, which may not be realistic in practice. The authors in [ 16 ] proposed an algorithm named Energy-Efficient Task Scheduling in Fog Computing based on Particle Swarm Optimization (EETSPSO). The results demonstrated that EETSPSO outperforms existing approaches in terms of makespan, energy consumption, and execution time. However, as fog nodes are resource-constrained, the algorithm’s performance may vary across different real-world scenarios with varying resource limitations. I-B Deadline constrains The task scheduling approach proposed in [ 7 ] incorporates two semi-greedy algorithms: Priority-aware Semi-Greedy (PSG) and PSG with Multi-start Procedure (PSG-M). These algorithms aim to optimize energy consumption and minimize deadline violations for IoT tasks by effectively mapping them to fog nodes while ensuring compliance with Quality of Service (QoS) requirements. Notably, the PSG-M algorithm demonstrates a significant reduction in deadline violation time. However, the performance of these algorithms may vary under different real-world scenarios characterized by diverse resource constraints. In [ 17 ] , Louail et al. introduced a dynamic task scheduling algorithm that considered both deadline constraints and task frequency to improve the efficiency of fog nodes in smart factories. By considering task frequency, the algorithm optimizes the utilization of available resources, thereby reducing idle time and enhancing overall efficiency. However, the algorithm was complex to implement and required significant computational resources, which could be a challenge in resource-constrained environments. I-C Latency Reduction Authors in [ 18 ] proposed an Energy-Efficient Internet of Medical Things to Fog Interoperability of Task Scheduling (EEIoMT) framework. This framework aims to schedule tasks efficiently by ensuring that critical tasks are executed