Constrained Multi-Objective Optimization for UAV Energy Consumption and User Satisfaction in UAV-Assisted Mobile Edge Computing

The energy consumption for unmanned aerial vehicles (UAV) and the user satisfaction in the UAV-assisted mobile edge computing (MEC) has been optimized considering the UAV supply side and the user demand side simultaneously. The user priority level in the UAV-assisted MEC model has been introduced to...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on vehicular technology pp. 1 - 16
Main Authors: Chai, Xuzhao, He, Cuicui, Qu, Boyang, Yan, Li, Li, Chao, Qiao, Baihao, Liang, Jing, Wen, Pengwei, Li, Zhao, Liu, Ping, Suganthan, Ponnuthurai Nagaratnam
Format: Journal Article
Language:English
Published: IEEE 2025
Subjects:
Constrained multi-objective optimization
Inte-grated Constraint strategy
NSGA-II
UAV energy consumption
User satisfaction
ISSN:0018-9545, 1939-9359
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The energy consumption for unmanned aerial vehicles (UAV) and the user satisfaction in the UAV-assisted mobile edge computing (MEC) has been optimized considering the UAV supply side and the user demand side simultaneously. The user priority level in the UAV-assisted MEC model has been introduced to match the real scenario. The UAV energy consumption has been modeled, as well as the user satisfaction. To solve the optimization problem effectively, the tasks are scheduled for the UAVs referring to the user priority. Subsequently, a constrained multi-objective evolutionary algorithm based on Non-dominated Sorting Genetic Algorithm II (NSGA-II) has been designed, merging the integrated constraint strategy, the adaptive genetic operator strategy, and the Q-learning-based selection mechanism. The integrated constraint strategy is adopted to comprehensively find the better solutions based on the scoring mechanism; the adaptive genetic operator strategy is used to adjust dynamically the crossover and mutation parameters; the Q-learning-based selection mechanism is to record individual experience and feedback for adaptively learning. In order to verify the performance of the algorithm, the nine experimental cases have been configurated. The IGD and HV indicators are used to characterize the performance compared with the other seven algorithms. The results show that the proposed algorithm has a great advantage in diversity and convergence, and can solve effectively the optimization for UAV energy consumption and user satisfaction in the UAV-assisted MEC.
ISSN:0018-9545
1939-9359
DOI:10.1109/TVT.2025.3624297