并行遗传算法在MRI永磁主磁体优化设计中的应用  

Parallel genetic algorithm and its application in optimization design of MRI main magnet assembly

作　　者：谢德馨[1] 何笑帆[1] 曾建斌[1] 

机构地区：[1]沈阳工业大学电气工程学院

出　　处：《沈阳工业大学学报》2008年第2期125-128,153,共5页Journal of Shenyang University of Technology

基　　金：国家自然科学基金资助项目(50377027)

摘　　要：为了减少电磁场逆问题求解的计算时间,提出了一种并行遗传算法,并采用并行虚拟机(PVM)方式,将该算法用于磁共振成像装置主磁体的极靴形状优化.研究表明:在合理确定并行计算的计算机台数和进程数的条件下,应用PVM实现的并行遗传算法可以有效地提高计算速度.数值实验验证了所提出的方法和所编制软件的实用性.由此可知,用PVM实现的并行遗传算法对于求解计算时间冗长的电磁场逆问题,是一种较好的方法.A parallel genetic algorithm was presented to reduce the computation time for electromagnetic field inversion problems.The method was applied to optimizing the pole shape of the main magnet in the magnetic resonance imaging（MRI） device,using the parallel virtual machine（PVM）.The research shows that the proposed method can shorten the computation time effectively if the number of parallel computers and processes are determined reasonably.Numerical experiments verified the validity of the proposed method and the compiled software.It can be seen that the parallel genetic algorithm realized by PVM is a good approach to solve the time-consuming inversion problems.

关 键 词：核磁共振成像 主磁体 优化设计 并行遗传算法 并行虚拟机 

分 类 号：TM351[电气工程—电机]