Dynamic Weight Fusion with Entropy Optimization for Enhanced Distributed Object Detection and Resource Allocation

The Dynamic Weight Fusion with Entropy Optimization algorithm introduces an innovative approach to enhancing distributed object detection and resource allocation in environments with limited resources, such as IoT networks and edge computing systems. Conventional object detection methods struggle to balance accuracy and resource efficiency when deployed on various devices with different computational strengths. The Entropy Weight Fusion algorithm tackles these challenges by integrating measures of uncertainty to dynamically modify the weight of each device’s contribution, prioritizing outputs with higher reliability while reducing the impact of uncertain detections. This strategy enables more effective aggregation of detection outputs, resulting in enhanced detection precision, minimized packet loss, and steady latency performance. Comprehensive evaluations highlight the algorithm’s effectiveness in achieving high utility scores and managing resources efficiently, making it a promising solution for real-time applications such as surveillance, autonomous systems, and other use cases requiring dependable distributed object detection. This work addresses the challenge of maintaining detection accuracy under limited resources in edge environments. By combining utility-based scoring with entropy-driven uncertainty handling, the proposed method enables adaptive, efficient object detection across heterogeneous devices.