TY - GEN
T1 - AI-Driven Task Offloading Prediction Using SAP-DE-Optimized GRU in 6G Edge Computing
AU - Gaurav, Akshat
AU - Pan, Shin Hung
AU - Arya, Varsha
AU - Hsu, Ching Hsien
AU - Chui, Kwok Tai
AU - Gupta, Brij B.
N1 - Publisher Copyright:
© 2026 IEEE.
PY - 2026
Y1 - 2026
N2 - Task offloading is a important issue in 6G edge computing, which required intelligent decision making. In this context, this paper proposed an AI-driven task offloading prediction model. The proposed model used differential evolution with self-adaptive Populations (SAP-DE) for feature selection and Gated Recurrent Unit (GRU) for classification. The proposed model achieves an accuracy of 80.57%, surpassing different baseline models. Compared to baseline models, it reduces FLOPs by 98.82% and total trainable parameters by 97.26%, ensuring lightweight deployment.
AB - Task offloading is a important issue in 6G edge computing, which required intelligent decision making. In this context, this paper proposed an AI-driven task offloading prediction model. The proposed model used differential evolution with self-adaptive Populations (SAP-DE) for feature selection and Gated Recurrent Unit (GRU) for classification. The proposed model achieves an accuracy of 80.57%, surpassing different baseline models. Compared to baseline models, it reduces FLOPs by 98.82% and total trainable parameters by 97.26%, ensuring lightweight deployment.
KW - 6G Edge Computing
KW - Differential Evolution with Self-Adaptive Populations (SAP˙DE)
KW - Gated Recurrent Unit (GRU)
KW - Task Offloading
UR - https://www.scopus.com/pages/publications/105037391544
U2 - 10.1109/ICCE67443.2026.11449888
DO - 10.1109/ICCE67443.2026.11449888
M3 - Conference contribution
AN - SCOPUS:105037391544
T3 - Digest of Technical Papers - IEEE International Conference on Consumer Electronics
BT - 2026 IEEE International Conference on Consumer Electronics, ICCE 2026
T2 - 2026 IEEE International Conference on Consumer Electronics, ICCE 2026
Y2 - 3 February 2026 through 5 February 2026
ER -