南京邮电大学学报（自然科学版）2026，Vol.46Issue(2)：75-83,9.DOI:10.14132/j.cnki.1673-5439.2026.02.009
基于差分进化搜索的边边协同计算任务卸载算法
A differential evolutionary search-based computing task offloading algorithm for edge-to-edge collaboration
摘要
Abstract
Edge computing power network(EdgeCPN),as a new computing paradigm,flexibly sched-ules fragmented computing resources in CPN to meet different task requirements,enabling efficient computing-task offloading for large-scale terminal scenarios.This paper proposes a mobile device comput-ing task offloading model based on edge-to-edge collaboration.This model divides task offloading in EdgeCPN into two phases:the construction of the edge computing resource pool and the allocation of end-edge resources.We propose a differential evolutionary search scheme based on the optimal-cost comput-ing resource pool to minimize the total delay of task offloading.First,a subset of the optimal cost comput-ing resource pool is selected based on the user's budget.Second,leveraging the available computing re-sources of the pool,the differential evolution algorithm is used to jointly optimize the task offloading deci-sions of the mobile devices with the objective of minimizing total delay,thereby matching each terminal's computing task to an appropriate server.Simulation results show that this scheme significantly improves the efficiency and stability of computing resource scheduling performance in EdgeCPN.关键词
算力网络/任务卸载/资源分配/边缘服务器协同Key words
computing power network(CPN)/task offloading/resource allocation/edge server collabo-ration分类
信息技术与安全科学引用本文复制引用
王琴,马雪晴,杨惠茗,朱洪波..基于差分进化搜索的边边协同计算任务卸载算法[J].南京邮电大学学报（自然科学版）,2026,46(2):75-83,9.基金项目
江苏省重点研发计划项目(BE2022068,BE2022068-2)、国家电网有限公司总部科技项目(521001250031-239-TD)、国家自然科学基金重大研究计划集成项目(92367302)和江苏省高等学校基础科学(自然科学)研究重大项目(24KJA510008)资助项目 （BE2022068,BE2022068-2）
