一种基于输运理论的多目标演化算法

提出了一种根据输运理论中的粒子输运方程、相空间能量定律和熵增法则构造的一种能够准确、高效地求解多目标优化问题的多目标演化算法(MOPEA).由于该算法使用了粒子系统从非平衡达到平衡的理论来定义求解多目标问题的Rank函数和Niche适应值函数,使得种群中的所有个体都有机会参与演化操作,以达到快速、均匀地求出多目标优化问题的Pareto最优解.数据实验显示,利用该算法求解多目标优化问题不仅能够使算法快速地收敛到全局Pareto前沿,同时由于该算法要求所有的粒子都要参与杂交和变异等演化操作,从而避免问题早熟现象的出现,并通过与传统演化算法的性能指标分析比较说明,使用该算法求解多目标优化问题具有明显的优越性.

A MOP Evolutionary Algorithm Based on Transportation Theory

In this paper a Multi-objective Optimization Problems Evolutionary Algorithm,MOPEA,for solving multi-objective optimization problems precisely and efficiently is presented according to the equation of particle transportation and the principle of energy decreasing and the law of entropy increasing in phase space of particles based on transportation theory.In the algorithm,the theory of particle system changing from non-equilibrium to equilibrium is used to define the Rank function and Niche function for solving multi-objective problems,all the individuals in the population have chance to participate the evolving operation to solve the Pareto optimal solutions of the multi-objective problems fast and evenly.The experiments show that this algorithm can not only converge to global Pareto optimal front fast and precisely,but also can avoid premature phenomenon of multi-objective problems because the algorithm requires all the particles in the phase space to cross and mutate simultaneously.Through analyzing the performance indices of evolutionary algorithms it illustrates that this algorithm have more advantages than traditional evolutionary algorithms.