论文部分内容阅读
提出了一种连续驻留位置和驻留时间的近距离放疗剂量优化算法。将驻留时间t看成是驻留位置x的一个连续函数,用积分方式计算每个参考剂量点的剂量。再用计算剂量和目标剂量值之间的差别为目标函数,用遗传算法和POWELL算法混合求解最优的曲线参数。曲线可根据具体情况进行分段。得到曲线后再根据积分的数值逼近方法原理将其离散化,得到最终的驻留位置和驻留时间。实验的结果表明,算法不仅避免了负的驻留时间问题,还让相邻驻留位置的驻留时间比较平滑。在最后的离散化过程中,还可以得到不同的驻留位置和驻留时间结果,使计划具有更好的灵活性。
A dose optimization algorithm of brachytherapy with continuous dwell position and dwell time is proposed. The dwell time t is considered as a continuous function of the resident position x, and the dose for each reference dose point is calculated by integration. Then the difference between the calculated dose and the target dose is taken as the objective function, and the optimal curve parameters are solved by genetic algorithm and POWELL algorithm. Curves can be segmented according to the specific situation. After the curve is obtained, it is discretized according to the principle of numerical approximation of the integral to obtain the final residence position and dwell time. The experimental results show that the algorithm not only avoids the problem of negative dwell time, but also makes the dwell time of adjacent dwell positions relatively smooth. During the final discretization, different dwell positions and dwell time results are also available, giving the plan greater flexibility.