Edge Computing Offloading via Lyapunov Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In mobile edge computing networks, existing technologies face challenges in efficiently managing resource allocation and offloading decisions due to user mobility, leading to increased energy consumption, task delay, and migration costs, especially with the integration of communication and sensing functions.
Innovation Solution
A method for offloading decision and resource allocation based on integration of communication, sensing, and computing, which involves obtaining user location, channel status, and task data volume, and performing initial and swap matching to determine base station selection, channel allocation, and task offloading ratio, while considering user mobility and sensing performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If user mobility is considered in resource allocation, then system adaptability is improved, but computational complexity increases
Solution Approach 1:
The patent segments the resource allocation problem into two distinct phases: an initial matching phase that establishes base assignments, and a swap matching phase that refines allocations. This segmentation allows the system to handle user mobility adaptably while keeping computational complexity manageable by breaking down the complex optimization into smaller, more tractable sub-problems.
Solution Approach 2:
The patent performs preliminary matching in the initial matching phase, establishing base resource allocations before considering swap operations. This preliminary action provides a starting point that reduces the search space for subsequent optimization, thereby improving adaptability to user mobility while controlling computational complexity.
2Use of energy by moving object
If task offloading is optimized, then energy consumption is reduced, but task delay may increase
Solution Approach 1:
The patent applies partial offloading where only a portion of user tasks are offloaded to edge servers rather than complete offloading. This allows the system to reduce energy consumption by leveraging edge computing resources while keeping critical or time-sensitive tasks local, thus balancing energy savings with task delay requirements.
Solution Approach 2:
The patent dynamically adjusts the offloading ratio parameter based on system conditions, user mobility, and task characteristics. By changing this parameter adaptively, the system optimizes the trade-off between energy consumption and task delay, offloading more when energy savings are prioritized and less when delay constraints are tighter.
Data Source
AI summary
A method for offloading decision and resource allocation based on integration of communication, sensing and computing is provided. A system cost function is defined by considering energy consumption, time delay, and migration costs. Under the constraints of user sensing failure rate and maximum task completion time delay, a long-term average cost minimization problem is established. Based on Lyapunov optimization theory, a virtual queue is established to evaluate the user sensing performance. By using Lyapunov drift-plus-penalty function, the long-term stochastic optimization problem is transformed into a deterministic optimization problem with a single time slot. The transformed problem is divided into inner and outer layers. The inner layer obtains the optimal task offloading ratio for each user base station selection and channel allocation through theoretical derivation. The outer layer determines the user base station selection and channel allocation based on the results of solving the inner layer problem through matching theory.