遗传算法优化放射治疗计划射野权重的收敛性分析

Convergence analysis of genetic algorithm for beam weights optimization in radiotherapy

随着三维适形和调强放疗技术的发展,放射治疗计划射野参数的优化问题受到了广泛关注。遗传算法作为一种全局优化技术,由于其不要求目标函数具有连续性、导数存在、线性等条件的假设,具有搜索速度快、搜索空间广的特点,可以用来解决很多优化问题,因此,在放射治疗计划射野参数优化中得到了越来越多的应用。对不同遗传参数条件下的遗传算法收敛性进行了分析,给出了分析结果并进行了讨论。

With the development of 3D CRT and IMRT, the optimization of beam parameters in external beam radiotherapy has gained wide interests. As one of the technologies for global optimization problems, Genetic Algorithm （GA） is always available no matter whether the objective function is continuous, differentiable and linear or not. Because of its high speed and vast search space, GA optimization is becoming more and more popular in optimization of beam parameters. In this paper, the convergence of GA with different genetic parameters is analyzed and the result is discussed.