摘要
基于分解的多目标进化算法MOEA/D(Multi-objective Evolutionary Algorithm Based on Decomposition)具有收敛速度快、分布性好等特点,但其在非凸函数上的性能有待提高。鉴于量子进化算法在多峰值函数上的优良性能,将MOEA/D与量子进化算法相结合,提出基于分解的多目标量子差分进化算法QD-MOEA/D(Quantum Differential Multi-objective Evolutionary Algorithm Based on Decomposition)。QD-MOEA/D的量子染色体采用实数编码,节省存储空间,加快运算速度。为了加快算法收敛速度并提高算法探测能力,量子染色体采取差分进化,其变异方式为量子非门。在多个标准测试函数的实验结果表明,该算法改进了MOEA/D在非凸函数上的收敛性和分布性。
M u lti-o b je c tiv e evo lu tiona ry a lgo rithm based on decom position ( M O E A /D ) is featured by hig h convergence rate and goodd is trib u tio n . H ow ever, its perform ance in non-convex fu n ctio n s is not good enough. In view o f the e xce lle nt properties o f quantum evolutionarya lg o rith m in m u lti-p e a k fu n c tio n s , we com bined M O E A /D w ith Q E A and proposed the decom position-based quantum d iffe re n tia l m u ltiob je ctive evo lu tiona ry a lgo rithm ( Q D - M O E A / D ) . The quantum chromosome o f Q D -M O E A /D adopts real nu m ber in e n coding , th is savesm em ory space and accelerates operation speed. In order to speed up convergence speed and im prove de tection a b ility o f a lg o rith m , thequantum chromosome adopts d iffe re n tia l e v o lu tio n , and its m u tation way is the quantum non-gate. Results o f experim ents on several standardtest fu n ctio n s showed th a t the a lgo rithm im proved the convergence and the d is trib u tio n o f M O E A /D in non-convex fun ctions.
作者
常新功
刘文娟
吕亚丽
Chang Xingong;Liu Wenjuan;Lu Yali(Faculty of Information and Management, Shanxi University of Finance and Economics, Taiyuan 030031 , Shanxi, China)
出处
《计算机应用与软件》
CSCD
2016年第8期277-282,共6页
Computer Applications and Software
基金
山西省自然科学基金项目(20130110164
2014011022-2)
山西省高校科技创新项目(2013124)
关键词
MOEA/D
量子计算
差分进化
实数编码
M O E A /D
Quantum computation
Differential evolution
Real-encoding
