双种群遗传算法优化射野方向及权重的研究  

A Study on IMRT Beam Weight and Orientation Optimization based on Genetic Algorithm with Two Populations

作　　者：曹明勇[1] 陈超敏[1] 周凌宏[1] 童蕾[2] 

机构地区：[1]南方医科大学生物医学工程学院,广州510515 [2]广东机电职业技术学院,广州510515

出　　处：《医疗卫生装备》2008年第10期4-6,23,共4页Chinese Medical Equipment Journal

基　　金：广东省自然科学基金(06024583);广东省科技计划项目(2007B01040056)

摘　　要：目的:研究双种群遗传算法在同时优化射野方向和射野权重中的应用。方法:建立基于三维光子笔射束的IMRT剂量计算模型,用VisualC#.Net编写剂量计算并实现双种群遗传算法优化射野方向和权重的算法,分析优化结果。结果:在射野较少的情况下射野方向的微小变化可以使靶区剂量分布更加适形,且高剂量区域更大,用双种群遗传算法同时优化射野方向和权重,能够在一个临床可接受的计算时间内得到较高适形度的剂量分布。结论:在射野较小的情况下,射野方向的改变对剂量的分布有很大的影响。双种群遗传算法是一种有效的随机全局优化方法,它的并行性、全局性、鲁棒性使得方向和权重两个具有复杂耦合关系的量能够同时优化。双种群遗传算法在IMRT计划优化中具有广阔的应用前景。Objective To study the application of the genetic algorithm with two populations in optimization of both IMRT beam weight and orientation.Methods The genetic algorithm with two populations was used to optimize IMRT beam weight and orientations with three dimensional pencil beam dose calculation model. The algorithm was implemented in Visual c#. Net. And the results were analyzed. Results The modest change of the beam orientation resulted in more conformal dose distribution in target volume, and large high dose area was included. Further more, genetic algorithm with two populations gave such a clinically acceptable computation time. Conclusion Little change had distinct influence on dose distribution when there were few beam orientations. The genetic algorithm with two populations was an effective and global optimization method because of its parallel, robust and global. It provided satisfying results when used to optimize IMRT beam weight and orientations .

关 键 词：双种群遗传算法 射野方向 经验约束 调强放疗 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]