CAJ | 학술논문

针对以往文献中高斯变异差分进化算法变异方式的导向性过于单一，不利于算法种群结构多样性保持，种群进化信息吸取过于单调的弱点，设计了一种差异化高斯双导向差分进化算法。采用向量图分析方法设计了一种新的高斯双导向变异方式，该变异方式能够兼顾全局进化、局部进化及个体进化信息，以当前全局最优值和个体历史最优值作为个体进化的两个不同导向，从而使变异后的个体能够吸收更多的种群进化有利信息并加以利用，并且根据种群个体进化差异化程度选取合适的变异方式。对差异化高斯双导向差分进化算法的性能进行计算机仿真设计验证，并对算法的物流配送优化问题进行研究。
침대이왕문헌중고사변이차분진화산법변이방식적도향성과우단일，불리우산법충군결구다양성보지，충군진화신식흡취과우단조적약점，설계료일충차이화고사쌍도향차분진화산법。채용향량도분석방법설계료일충신적고사쌍도향변이방식，해변이방식능구겸고전국진화、국부진화급개체진화신식，이당전전국최우치화개체역사최우치작위개체진화적량개불동도향，종이사변이후적개체능구흡수경다적충군진화유리신식병가이이용，병차근거충군개체진화차이화정도선취합괄적변이방식。대차이화고사쌍도향차분진화산법적성능진행계산궤방진설계험증，병대산법적물류배송우화문제진행연구。
According to the differential evolution improvement oriented mutation by Gauss sampling too single in the previous literature,was not conducive to the algorithm to maintain the population diversity structure,population evolution information learned too monotonous weaknesses,this paper designed a kind of difference Gauss double guide differential evolution algo-rithm.Firstly,using the vector analysis method to design a new Gauss dual oriented variation,which could take into account the overall evolution,individual local evolution and evolutionary information,took the current global optimal value and the optimal value of individual history as two different guiding the individual evolution,so that the mutated individual could absorb more advantage information of population evolutionary and made use,and chose the right mutation operation according to the individ-ual.Then,verify the performance of differentiation Gauss double orientation differential evolution by computer simulation,and applies it to the logistics distribution optimization.