The Application of Intelligent Algorithms in Energy Management of Smart Cities

It is difficult to balance energy supply and demand in smart city energy management, and the traditional scheduling strategy has poor real-time performance. This paper proposes an intelligent energy management system that integrates deep learning and reinforcement learning: first, a spatiotemporal graph convolutional network is constructed to capture the topological characteristics of city-level energy flows, and an improved LSTM network is used to integrate 12-dimensional time series data such as meteorology and social activities to achieve 96-hour load forecasting. Secondly, a dual-depth Q network dynamic scheduling engine is designed, and a decision model is established through collaborative optimization of value network and target network. Then a multi-objective particle swarm optimization algorithm is developed to solve the Pareto boundary, and blockchain technology is combined to realize trusted transactions between microgrids. In the Suzhou Industrial Park demonstration project, the system increased the photovoltaic absorption rate to 99.2% and shortened the response time to 8.7 seconds, verifying that intelligent algorithms can improve energy management efficiency and provide a reliable technical path for the low-carbon transformation of smart cities.