Geometrized Task Scheduling and Adaptive Resource Allocation for Large-Scale Edge Computing in Smart Cities

With the widespread adoption of Internet of Things (IoT) devices, the number of smart nodes in smart cities is experiencing explosive growth, posing unprecedented challenges to edge computing. On one hand, the deployment of large-scale distributed edge nodes and the massive demand for task processing make it difficult for traditional optimization theory-based scheduling methods to cope with the resulting surge in computational complexity; on the other hand, the randomness and high concurrency of task requests also place higher demands on resource allocation. To address these issues, there is an urgent need to research efficient large-scale edge computing task scheduling and resource allocation strategies tailored to the specific scenarios of smart cities. This has become a key technical challenge that urgently needs to be overcome.

In response to the challenges above, the research team has innovatively proposed a task scheduling framework for large-scale edge computing in smart cities. This framework transforms the large-scale task scheduling problem into a geometric partitioning problem, achieving efficient task allocation by introducing stream clustering techniques and weighted Voronoi diagrams, which greatly reduces the solution complexity. At the same time, the team also designed a “Tetris-like” task offloading evaluation mechanism and proposed a dynamic resource allocation strategy based on adaptive sliding windows, effectively solving the resource allocation problem under random and high-concurrency task requests. Taking intelligent transportation monitoring as an example, the framework can efficiently process real-time monitoring tasks from a large number of sensors, ensuring quality of service while significantly reducing computational complexity and improving system performance. Experimental results show that this method significantly reduces the task deadline violation rate, achieving a performance improvement of more than 20 times compared to existing methods in various test scenarios. The research findings provide important theoretical and technical support for large-scale edge computing applications in future complex large-scale systems, such as smart cities and unmanned offshore central platforms.

The related research findings were published in the IEEE Internet of Things Journal under the title “Geometrized Task Scheduling and Adaptive Resource Allocation for Large-Scale Edge Computing in Smart Cities”. The first author of the paper is Chen Yang, a doctoral student from the Institute of Ocean Engineering at Tsinghua University Shenzhen International Graduate School (Tsinghua SIGS). The corresponding author is Assistant Professor Zheng-Ru Ren from Tsinghua SIGS. Co-authors include Associate Professor Yuemin Ding from the University of Navarra, Spain, and Associate Professor Zhenzhong Hu from Tsinghua SIGS. This project was funded by the National Key R&D Program of China (Ministry of Science and Technology).

[Download Paper]