有限元结合并行梯度遗传算法及其应用

摘要

This paper introduces a new kind of improved genetic algorithm-parallel gradient genetic algorithm. The principle is that the parallel computing and the gradient algorithm are combined to genetic algorithm. This kind of method can make full use of computer multi-core CPU, but also can accelerate the convergence time of genetic algorithm to improve computational efficiency. The algorithm is applied to the finite element analysis of the permanent magent linear synchronous motor( PMLSM). Analysising the magnetic field and the horizontal thrust of PMLSM by programming Ansoft 3D macro functions also opti-mizating and improving the horizontal thrust of unit volume of PMLSM. The results showed that the time of this algorithm is short and the the horizontal thrust of unit volume of PMLSM has been significantly improved.%文章介绍一种新型的改进遗传算法-并行梯度遗传算法.其原理是把并行计算和梯度算法结合到遗传算法中.这个种方法既可以充分利用计算机多核CPU,又可以加速遗传算法的收敛时间,提高计算效率.将该算法应用于永磁直线同步电机( PMLSM)有限元分析中.通过编程实现Ansoft3D宏功能,对PMLSM进行磁场和水平推力分析,并对PMLSM单位体积水平推力进行优化,达到提高PMLSM单位体积推力的目的.结果表明:该算法优化时间短,并且电机的水平推力得到显著提高.