基于粒子群算法的γ能谱重叠峰解析

Decomposing the overlapping peaks inγspectrum based on particle swarm algorithm

针对传统γ能谱分析中最小二乘法中的迭代算法对初始峰参量要求较严格,并且容易陷入局部最优等问题,提出了基于粒子群算法的γ能谱重叠峰解析方法,并且给出了重叠峰解析原理以及相应的算法.模拟和实测γ能谱重叠峰的解析结果均表明:在重叠峰的分离度较低的情况下,该方法能够得到较高的解析精度,并且具有较强的解析能力.另外,该方法还具有参量少、对初始参量要求宽松和算法易于实现等优点,而且收敛于全局最优解,因而是有效的γ能谱重叠峰解析方法.

To overcome the disadvantages such as requiring strict initial parameters,easily falling into the local optimum,of the iterative algorithm used in the traditionalγspectrum analysis,a particle swarm algorithm based the method for decomposing the overlapping peaks ofγspectrum was proposed,and the basic principle and corresponding algorithm were given.The results of decomposing the overlapping peaks for the simulated and measuredγspectra showed that the method could achieve high analytical accuracy even under the situation of the low separation degree of overlapping peaks,while having strong decomposing ability.Meanwhile,the algorithm has the advantages of fewer parameters,relaxed requirements on initial parameters,and easy implementation.And converging to the global optimal solution made it an effective method for decomposing the overlapping peaks inγspectrum.