Service Demand Forecasting and Service Composition Optimization Method in Cloud Manufacturing Environment

In cloud manufacturing environments, fluctuations in customer service demands often impose constraints on the allocation efficiency of cloud service resources for machine tool equipment. To enhance the adaptability of service composition optimization methods to dynamic service demands and improve optimization efficiency, this paper proposes a service composition optimization framework integrating LSTM neural networks with genetic algorithms. The framework operates through three key phases: First, historical service demand data is utilized to perform demand forecasting via an LSTM model. Second, a discrete-time sequence service composition optimization model is established based on the predicted data. Finally, genetic algorithms are employed to solve the optimization problem. Simulation experimental results demonstrate that the proposed framework effectively captures fluctuation characteristics in service demand data and delivers robust optimization performance for service composition selection under demand volatility.