Event-Driven Swarm Optimization for Energy–Latency Tradeoffs in Asynchronous Edge Computing Systems

Abstract

Edge computing platforms increasingly host latency-sensitive and energy-constrained applications that operate over distributed and heterogeneous resources. As workloads become more dynamic and spatially distributed, static or centrally orchestrated optimization strategies face difficulty in maintaining efficient energy–latency tradeoffs under stringent scalability and responsiveness constraints. Asynchronous execution models at the edge introduce further challenges due to irregular task arrivals, heterogeneous hardware, and nonuniform communication delays that disrupt classical synchronous optimization loops. This work examines an event-driven swarm-based optimization approach that aims to coordinate edge nodes under asynchronous conditions while explicitly balancing energy consumption and end-to-end service latency. The proposed methodology replaces periodic global coordination with local, event-triggered updates that allow each node to react only when relevant system states change significantly, thereby limiting redundant computation and communication overheads. Swarm-inspired search mechanisms are employed to explore allocation and offloading decisions in a distributed fashion, while the energy–latency compromise is captured through a tunable objective that can adapt to diverse application preferences and operational regimes. A linear modeling framework is used to represent energy and latency components, which supports efficient evaluation of candidate configurations during the swarm search process. Simulation-based analysis illustrates how the event-driven swarm mechanism adapts to variation in workload intensities, communication delays, and device energy profiles, and how different parameter settings shape the resulting energy–latency tradeoff surfaces. The results suggest that asynchronous and event-driven swarm coordination can be a practical design option for large-scale edge computing deployments that operate under heterogeneous and time-varying conditions.