Computation Offloading Based on Unscented Transform Biogeography-based Optimization in Mobile Edge Computing

• Published in: 2024 39th Youth Academic Annual Conference of Chinese Association of Automation (YAC) pp. 98 - 103
• Main Authors: Pu, Guiyang, Sun, Hongchun, Dai, Jun, Na, Xiaodong, Jiang, Mingfeng, Li, Yanshan
• Format: Conference Proceeding
• Language: English
• Published: IEEE 07.06.2024
• Subjects: biogeography-based optimization; computation offloading; Cost function; Costs; Decision making; Energy consumption; mobile edge computing; Multi-access edge computing; Performance evaluation; Transforms; unscented transform
• ISSN: 2837-8601
• DOI: 10.1109/YAC63405.2024.10598691

Computation offloading, as one of the means to reduce latency and energy consumption in mobile edge computing (MEC), can significantly reduce communication costs and enhance resource utilization. In response to compute-intensive and latency-sensitive application scenarios, an unscented transform biogeography-based optimization (UTBBO) algorithm is proposed to the facilitate computation offloading of MEC. Building upon classical migration and mutation strategies, UTBBO utilizes a guiding strategy with unscented transform to finely develop elite individuals within the visible range of the search domain, thereby avoiding the algorithm from falling into local optima. Furthermore, a computation offloading system encompassing cloud, edge, and terminal devices is established, and the offloading tasks are transformed into the optimization process of the cost function using UTBBO, where the cost function is a weighted sum of latency and energy consumption. Experiments demonstrate that the proposed UTBBO algorithm can effectively reduce latency and energy consumption in computation offloading tasks across multiple scenarios in edge computing, providing more rational offloading decisions.