TL;DR
A distributed Multi-stage Adaptive Deferred Acceptance (MA-DA) algorithm is proposed that enables a stable and Pareto-optimal assignment of tasks to edge computing nodes (ECNs) and determines a reasonable task execution sequence and ensures the prioritized completion of delay-sensitive tasks.
This work aims to design an efficient framework by incorporating a novel hybrid metaheuristic algorithm that combines Draco Lizard Optimization (DLO) and Sand Cat Optimization (SCO) for optimal task offloading and resource allocation for IoT applications.
: Cloud Computing (CC) is one of the widely used technologies due to its advanced features such as pay-per-use, scalability, and flexibility. The primary objective of CC is to allow users to access and purchase cloud services that are on demand through internet-based applications. Efficient load-balancing in the cloud faces challenges of high-dimensional state spaces and scalability with increasing tasks. To solve this problem, the Masterpiece Optimization Algorithm (MOA) with a priority constraint is employed for load-balancing according to the tasks efficiently. The MOA is integrated with a priority-based cost function to enhance the task scheduling process by introducing a multi-dimensional approach for load balancing. The priority-based framework helps the scheduler to dynamically recalibrate workloads. The experimental results achieve a total energy consumption of 39.8 W and an average CPU resource utilization of 99.54%, which is better than the existing algorithms, such as the hybrid Particle Swarm Grey Wolf Optimization (PSGWO) algorithm.
The backbone of modern digital machines is cloud computing′s on-demand resources, which scale to increase throughput. However, the field is increasingly complex from the point of view of heterogeneous cloud environments: requiring paradigms for virtual machines, containers and serverless computing nodes that put further challenges into ensuring an optimal performance/cost/energy balance. In traditional resource allocation and load balancing strategies, these goals are usually treated separately. As a result, there may be inadequate system utilization by either party involved in the transaction; ends responded with slow feedback time while intermediary services performed worse than predicted; or Service Level Agreements weren't satisfied. To help meet these challenges, this paper proposes a Hybrid Framework for Joint Optimization of Resource Allocation and Load Balancing that spans two layers in heterogeneous cloud computing systems. The framework integrates Dynamic Voltage and Frequency Scaling (DVFS) with migration-aware task assignment, learning-aided metaheuristic optimization to dynamically adjust to fluctuations in load while minimizing power consumption and operational costs. The optimization engine blends predictive workload modelling implemented via machine learning for proactive decision-making and hybrid solution strategies. Among its multiple objectives, this mechanism offers the ability to jointly minimize SLA violations, response time as well power usage. Experiments using a larger CloudSim and actual workload traces (extended CloudSim) show that the proposed framework obtains 25-30% energy savings compared with ordinary methods, significantly reduces p95 latency and also achieves a relatively better Quality Of Service. These results suggest the framework may provide a sustainable schedule which responds to outside influences in real time, making it suitable for future intelligent cloud system with large-scale organization.
The rapid advancement of edge computing has transformed the distributed computing paradigm. Edge computing enables storage and computation to perform at the network edge. For optimized performance, offloading has a crucial role that enhances the system efficacy by improving the quality of service (QoS) parameters. Therefore, this work proposes an Artificial Rabbit Optimization (ARO)-based framework for efficient task offloading-based allocation of resources in edge computing environments. The proposed technique improves the performance of Internet of Things (IoT) applications by using an intelligent task execution strategy that diminishes energy consumption, delay, and cost. A multi-objective function is framed that considers the above performance metrics and optimizes performance subject to delay and power constraints. The experimental results demonstrate that the proposed framework outperforms the benchmark approaches, reducing delay by up to 44.06%, energy consumption by up to $\mathbf{3 8. 9 6 \%}$, and cost by up to $\mathbf{2 8. 4 8} \boldsymbol{\%}$.
: Mobile edge computing (MEC) is an effective paradigm for supporting latency-sensitive and computation-intensive intelligent applications. However, in dynamic mobile-edge network scenarios, mobile terminals experience time-varying wireless links due to mobility. Tasks may also arrive unpredictably, while multiple terminals compete for limited edge resources. As a result, MEC systems may suffer from service congestion and unbalanced resource utilization, which increases end-to-end latency and energy consumption. This paper investigates cooperative task offloading in dynamic MEC networks. The considered system comprises one macro base station and multiple small base stations equipped with edge-computing resources. In each time slot, each mobile terminal selects a service option, determines the task offloading ratio, and chooses its transmit power for task uploading. This sequential decision process is formulated as a multi-agent problem with continuous action spaces. Under the centralized training and decentralized execution (CTDE) framework, the problem is further modeled as a decentralized partially observable Markov decision process (Dec-POMDP). Standard multi-agent soft actor-critic (MASAC) is not fully suitable for this problem. Its original action model does not handle bounded continuous actions well. Its exploration strength may also be unsuitable at different training stages. Frequent policy updates can further make training unstable when critic estimates are inaccurate. To address these issues, this paper develops an adaptive Beta-policy and delayed-update multi-agent soft actor-critic method, abbreviated as ABDMASAC. This method uses a Beta policy to model bounded actions. It adjusts the entropy coefficient during training and delays policy updates to reduce training oscillations. Experimental results show that, under a unified training budget and a consistent evaluation protocol, the proposed method achieves a better overall trade-off than the selected MASAC-backbone and on-policy MARL baselines under the considered simulation settings in terms of overall reward, average end-to-end latency, and average energy consumption. In the large-scale scenario, compared with MASAC, it improves the overall reward by 17.8%, reduces the average end-to-end latency by 18.0%, and lowers the average energy consumption by 11.4%.
Dynamic edge networks suffer from fluctuating bandwidth, latency, and edge-node load, which weaken conventional task offloading strategies under deadline and energy constraints. This study proposes an adaptive task offloading strategy based on dynamic network states. An online broad learning system predicts short-term bandwidth, transmission delay, and load from sliding-window observations. The predictions are embedded into a model predictive control framework for rolling optimization of the local-edge task allocation ratio, while confidence-interval-guided compensation adjusts decisions under prediction uncertainty. ADMM is used for distributed solution. NS-3/Python simulations show that the proposed method achieves a $93\%$ task success rate, 1.22 W average terminal energy consumption, 3.2 switches/s, and 45 ms average completion time, outperforming local execution, reactive offloading, and DQN-based offloading. The proposed strategy improves real-time reliability, energy efficiency, and robustness in highly dynamic edge computing environments.