中文摘要：

通过系统研究多目标粒子群算法,对于标准粒子群使用的线性惯性权重或常值惯性权重方法进行分析,发现粒子后期收敛速度的不足,针对这一问题,采用非线性递减指数函数的惯性权重取值方法,对粒子群速度更新公式进行分析研究,发现在算法迭代后期许多粒子速度停滞为零,易使粒子陷入局部最优,无法找到全局最优解,进而又提出了添加二次函数类速度扰动项的改进粒子群算法,该改进算法避免了粒子在迭代后期的停滞,使粒子在迭代后期仍具有较小的飞行速度,从而避免了粒子后期陷入局部最优。通过试验对比,改进后算法在收敛性和分布性能上均提高（30~50）%左右。

英文摘要：

Researched of multi-objective particle swarm optimization algorithm,and analyzed the method of linear or constant inertia weight in standard particle swarm. There was a fault at the end of searching,Which was low convergent speed. It adopted the exponential function of the nonlinear decreasing inertia weight value method to solve this problem. At the same time,analyzed the particle swarm velocity updating formula,most particles velocity of the later iteration algorithm stagnation was zero,which led to trap in local optimum easily,and was unable to find a global optimal solution. Aiming at the shortcoming,it puts forward the improved particle swarm algorithm to add quadratic function class of velocity perturbations.The improved algorithm avoided the iterative particle in later iteration,and the particles still had small flying speed,to avoid falling into local optimal particle later. Through comparative test,the improved algorithm in convergence and distribution performance was increased by（30~50）%.