Multi-criteria optimization framework for cloudlet computing in WMANs: integrating VL-WIDE and AHP for enhanced decision-making
As the demand for high-performance computing and low-latency applications escalates in wireless metropolitan area networks (WMANs), efficient task offloading and resource management have become increasingly critical. This article presents a framework for decision-maker satisfaction in WMAN cloudlet...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Published: |
John Wiley and Sons
2026
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/123161/ https://onlinelibrary.wiley.com/doi/10.1002/dac.70408 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | As the demand for high-performance computing and low-latency applications escalates in wireless metropolitan area networks (WMANs), efficient task offloading and resource management have become increasingly critical. This article presents a framework for decision-maker satisfaction in WMAN cloudlet computing optimization, addressing the challenge of balancing multiple objectives such as delay, energy consumption, and deployment cost in dynamic WMAN environments. Our method employs a variable-length multi-objective whale optimization integrated with differential evolution (VL-WIDE) to tackle the optimization problem, dynamically adjusting to the WMAN environment while optimizing delay, energy consumption (both user device and cloudlet), and cloudlet deployment cost. The analytical hierarchy process (AHP) is incorporated to balance the relative importance of these multiple criteria based on decision-maker preferences. Comprehensive time-series evaluations demonstrate that our framework significantly outperforms conventional systems and other benchmark algorithms, with VL-WIDE and AHP achieving the lowest root mean square error (RMSE) of 16.7, a 5.1% improvement over random selection. The proposed approach reduced execution and wireless delays by 40% (0.00015 s vs. 0.00025 s) compared to random selection, decreased mobile device energy consumption by 36.9% (118.69 Jol vs. 188.17 Jol), and demonstrated a 33.3% reduction in average renting rates compared to other algorithms. These results indicate that our proposed framework provides high satisfaction to decision-makers while significantly enhancing user experience in metropolitan areas, offering a robust solution to cloudlet-based computing optimization in WMANs and enabling higher efficiency and performance across multiple critical metrics. However, while the framework shows promising results, it is important to note that real-world implementation may face challenges not fully addressed in this study, such as network dynamics, scalability issues, and potential security concerns. |
|---|