Enhanced Multi-Objective Evolutionary Framework for Fluence Map Optimization in Intensity-Modulated Radiation Therapy

This paper presents an enhanced evolutionary framework for addressing the Fluence Map Optimization Problem (FMOP) within the context of Intensity-Modulated Radiation Therapy (IMRT). IMRT, as a critical technique in cancer treatment, requiring precise adjustments to radiation beams' intensity for optimal dose distribution within the patient's body. The proposed framework combines multi-objective evolutionary optimization with a gradient-based initialization strategy to achieve better convergence and diversity among the population of solutions. Extensive experiments on a clinical prostate case treatment plan demonstrate that the framework effectively optimizes the distribution of radiation intensity, resulting in improved protection of target structures and organs at risk. By leveraging a combination of evolutionary algorithms and gradient-based initialization, this study provides an innovative approach to solving the complex and large-scale FMOP, generating multiple clinical solutions at once.