非线性下降因子智能单粒子算法

Intelligent Single Particle Optimizer Based on Nonlinear Decreasing Factor Strategy

下降因子是基本智能单粒子算法中一个非常重要的参数,控制着算法由全局搜索转向局部搜索的速度,平衡算法的全局和局部搜索性能。基本的智能单粒子算法采用定值下降因子,不能随着计算的需要实时作出调整,对算法的性能造成了一定的不利影响。在基本的智能单粒子算法基础上,通过引入最大下降因子和控制因子这两个参数,提出一种基于非线性策略下降因子智能单粒子算法,成功地使得下降因子能够根据计算的需要而实时改变。函数测试和气动优化结果表明,改进算法能在相同的计算规模下获得比基本智能单粒子算法更好的解。

Decreasing factor is a very important parameter to the intelligent single particle optimizer since it balances the global searching ability and the local searching ability of the algorithm. Adopting a con- stant decreasing factor which should have been adjusted along with the calculation reduces the perform- ance of the conventional intelligent single particle optimizer. By introducing a max- decreasing factor and a controlling factor which succeed in making the decreasing factor change with the need of the intelligent single particle optimizer we propose a new algorithm called intelligent single particle optimizer based on nonlinear decreasing factor strategy. Function test and aerodynamic optimization experimental results have demonstrated that the improved algorithm performs much better than intelligent single particle optimizer on condition of the same number of function evaluations.