异步变化收缩因子粒子群算法心脏电流模型重构
摘要: 针对收缩因子的粒子群算法（CFPSO ）求解复杂问题时精度偏低，提出了异步变化收缩因子的粒子群算法（NL-CFPSO）。该算法利用异步变化的学习因子来产生收缩因子，改善粒子进化中的社会性和自身的学习性能，同时引入了“发散”和“收敛”操作，有效提高了粒子的收敛速度和精度。在该算法中，首先在给定的搜索空间上随机产生初始的粒子种群，在进化过程中采用异步变化的收缩因子，再根据判别函数来执行“发散”和“收敛”操作，使得粒子加速向全局最优的位置运动。将新算法和CFPSO在最新的6个测试函数进行对比，结果证明新算法比CFPSO算法具有更高的搜索精度和较低的时间复杂度。将该算法用于心脏单电流偶极子模型的反演计算仿真中，仿真结果证明该算法可以得到精确的模型参数，能够反映心脏模型的电磁现象，具有很高的实用价值。
On Reconstruction of Cardiac Current Model Based on Particle Swarm Optimization of Asynchronous Varied Constrict Coefficients
Keywords:
- 粒子群算法 /
- 收缩因子 /
- 异步变化 /
- 收敛精度
Abstract: Constrict Factor Particle Swarm Optimization (CFPSO) is low in accuracy ,and a new Particle Swarm Optimization based on asynchronous varied coefficients has been proposed in this paper .The algo-rithm based on CFPSO ,with asynchronous changing accelerating coefficients to produce the constrict fac-tor ,improve the sociality and self-studying of particles in evolutionary process ,and includes the operation known as “Diffusion” and“Attraction”in order to ameliorate the accurate level and convergence velocity . In this algorithm ,first of all ,produces random particles have been given with certain ranges .Then ,the a-synchronous constrict factor is used during the evolutionary procedure .Based on the discrimination func-tion ,the operation“Diffusion”or“Attraction”has been used ,which can evaluate the distance between the particles and global optima ,so the particles are moving toward to the global optima very rapidly .The new algorithm and CFPSO are both tested on novel six testing functions .The results indicate that the algo-rithm NL-CFPSO is more accurate than CFPSO ,and with a low time complexity .The algorithm is utilized to the calculating simulation of cardiac electromagnetic model .And the result can indicate that the accurate parameter of the model can be obtained by the algorithm ,w hich reflects the electromagnetic phenomena of cardiac model very accurately .It is high utilization .
下载: