Two-Stage Optimization Method for Beam Optical Systems Based on Multi-Objective Intelligent Optimization Algorithms

To achieve high-precision engineering design and improve the optimization efficiency of beam optical systems, this paper proposes a two-stage optimization method based on intelligent optimization algorithms. Firstly, the method employs a neural network to calibrate the transfer matrix of beamline components, followed by a two-stage strategy that integrates rapid transfer matrix optimization with precise particle tracking optimization to improve the system’s performance incrementally. The proposed method has been validated using a representative beamline case. Under constraints such as phase-space matching, the optimization results significantly improve key performance metrics, including transverse beam size and divergence. Compared to optimization approaches based solely on the particle tracking program, the optimization time is reduced from several days to a few hours, demonstrating a substantial enhancement in efficiency. By balancing simulation accuracy with computational cost, the method exhibits excellent generality and scalability in optimizing complex beam optics systems.