文献类型：期刊文献

英文题名：A Particle Swarm Optimization With Levy Flight for Service Caching and Task Offloading in Edge-Cloud Computing

作者：Gao, Tieliang[1];Tang, Qigui[1];Li, Jiao[1];Zhang, Yi[1];Li, Yiqiu[1];Zhang, Jingya[1]

通讯作者：Tang, QG[1]

机构：[1]Xinxiang Univ, Key Lab Data Anal & Financial Risk Predict, Xinxiang 453003, Henan, Peoples R China

第一机构：新乡学院

通讯机构：[1]corresponding author), Xinxiang Univ, Key Lab Data Anal & Financial Risk Predict, Xinxiang 453003, Henan, Peoples R China.|[11071]新乡学院;

年份：2022

卷号：10

起止页码：76636-76647

外文期刊名：IEEE ACCESS

收录：;WOS:【SCI-EXPANDED(收录号:WOS:000831057000001)】；

基金：This work was supported in part by the National Natural Science Foundation of China under Grant 61902021, Grant 61975187, and Grant 62072414; in part by the Key Science and Technology Program of Henan Province under Grant 222102210218, Grant 212102210096, and Grant 212102210104; and in part by the Soft Science Research Project of Henan Province under Grant 222400410137.

语种：英文

外文关键词：Task analysis; Cloud computing; Particle swarm optimization; Heuristic algorithms; Optimization; Internet of Things; Edge computing; Edge-cloud computing; particle swarm optimization; service caching; task offloading

摘要：Edge-cloud computing is an efficient approach to address the high latency issue in mobile cloud computing for service provisioning, by placing several computing resources close to end devices. To improve the user satisfaction and the resource efficiency, this paper focuses on the task offloading and service caching problem for providing services by edge-cloud computing. This paper formulates the problem as a constrained discrete optimization problem, and proposes a hybrid heuristic method based on Particle Swarm Optimization (PSO) to solve the problem in polynomial time. The proposed method, LMPSO, exploit PSO to solve the service caching problem. To avoid PSO trapping into local optimization, LMPSO adds levy flight movement for particle updating to improve the diversity of particle. Given the service caching solution, LMPSO uses a heuristic method with three stages for task offloading, where the first stage tries to make full use of cloud resources, the second stage uses edge resources for satisfying requirements of latency-sensitive tasks, and the last stage improves the overall performance of task executions by re-offloaded some tasks from the cloud to edges. Simulated experiment results show that LMPSO has upto 156% better user satisfaction, upto 57.9% higher resource efficiency, and upto 155% greater processing efficiency, in overall, compared with other seven heuristic and meta-heuristic methods.