Optimizing Social Resource Allocation via an IoT-Driven Predictive Electrical System

Abstract

Inefficient allocation of social resources for vulnerable urban populations presents a critical challenge. This study develops and validates an intelligent system for the predictive optimization of these resources, framed as a complex electrical and computational system. The system employs a hybrid Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) model for demand forecasting, utilizing data from a large-scale network of 5,320 Internet of Things (IoT) sensors. A Genetic Algorithm-Particle Swarm Optimization (GA-PSO) hybrid method, integrated with GIS, subsequently optimizes resource allocation and routing. This paper presents a novel contribution by being the first to systematically integrate this specific combination of predictive and optimization algorithms for social resource allocation, demonstrating significant performance gains over state-of-the-art methods. A new hierarchical system architecture is proposed to enhance scalability and real-time processing. Empirical validation involving 20 social work organizations demonstrated a demand prediction accuracy of 94.8%, a 96.2% reduction in service response time, and 99.2% system availability. The system also achieved high computational efficiency and low energy consumption, critical for sustainable deployment. This research delivers a validated and scalable engineering framework that significantly enhances the efficiency of urban social services, providing a replicable model for data-driven smart city applications at the intersection of electrical systems and computer science.