.png){kind=icon}
.png){kind=icon}
Return
Secure Computation Offloading Using Enhanced Genetic Algorithm for Ocean IoT
.png){kind=icon}
PRE
AI
DOI：10.1007/s10723-025-09797-3.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The network nodes in the Ocean Internet of Things exhibit strong heterogeneous characteristics, which introduce complex and high-dimensional constraints for the optimization of task offloading in ocean mobile edge computing. To prevent data from being intercepted by malicious attackers during computation offloading and to protect client privacy, an offloading framework named the Enhanced Genetic Algorithm for Secure Computation Offloading is proposed. The population initialization and crossover operators are optimized before the initialization phase to accelerate convergence and avoid the risk of excessive resource consumption due to cyclic scheduling. An elitism strategy is also introduced, along with dynamic mutation rate adjustment, to enhance the search efficiency and global optimization capability of the algorithm. The superiority of the proposed framework is validated through simulation experiments and a comprehensive comparison with other benchmark algorithms. Compared to the Simulated Annealing, Particle Swarm Optimization algorithms, the Min-Min algorithm, Basic Genetic Algorithm and Proximal Policy Optimization, the proposed scheduling algorithm demonstrates a significant enhancement in secure capacity during local computation, with improvements of 47.60%, 42.16%, 15%, 25% and 16.67%, respectively. Additionally, it achieves an increase in security during the computation offloading process by 16.54%, 16.22%, 14.29%, 12.56% and 10.02% respectively.
Keywords:
Ocean internet of things
Mobile edge computing
Secure computation offloading
Enhanced genetic algorithm
Privacy protection
Journal
IF:
2.9
Papers:
759
Citations:
1.2K