中文摘要：

针对调强放射治疗中基于等效均匀剂量线性目标函数的物理生物混合准则模型的不足，利用正则化理论构造了最大函数的平滑和凸正则化函数，改进了基于等效均匀剂量线性目标函数的混合准则放疗规划模型，解决了原模型限制优化算法寻优能力和难以确定梯度算法步长的问题。实验证明，该方法可以在保证相似靶区剂量覆盖特性的前提下，更好地保护危及器官，提高了放疗计划质量。

英文摘要：

To solve the problem existing in the physical-biological hybrid criteria model based on the generalized equivalentuniform dose linear objective function in conformal intensity- modulated radiation therapy, this paper presents asmooth and convex regularization of max function using regularization theory, and improves the hybrid criteria modelbased on the generalized equivalent uniform dose. When this improved model is applied in inverse planning optimizationprocess, problems(searching ability of optimization algorithm is restricted and it is hard time predicting how large a stepcan be made when using a gradient-based optimization algorithm)existing in original model are circumvented. Comparedto the original model, experiments show that the method using newly proposed model can get better protection to organsat risk while ensuring similar dose coverage for target and the quality of radiotherapy planning is improved.