基于动态迁移的ε约束生物地理学优化算法-论文.pdf

《基于动态迁移的ε约束生物地理学优化算法-论文.pdf》由会员上传分享，免费在线阅读，更多相关内容在应用文档-天天文库。

1、计算机研究与发展D()I：l0．7544“ssnl000—1239．20l4．20120666JournalofComputerResearchandDevelopment51(3)：580—589，2014基于动态迁移的s约束生物地理学优化算法毕晓君王珏李博李吉成(哈尔滨工程大学信息与通信1二程学院哈尔滨150001)(哈尔滨工业大学电子与信息丁程学院哈尔滨150001)(bixiaojun@hrbeu．edu．cn)An￡ConstrainedBiogeography‘BasedOptimizationwithDynamicMigrat

2、ionBiXiaojun。WangJLie，I．iBo，andLiJicheng(College0，Informationand(7ommunicationEngineering，HarbinEngineeringUniversity，Harbin150001)(Scbool『J／Electronic～andInformationEngineering，HarbinInstituteofTeehnology，Harbin150001)AbstractAnew￡constrainedbiogeography—basedoptimization

3、withdynamicmigration，eBB()一dm，isproposedtosolveconstrainedoptimizationproblems．Intheproposedalgorithm，the￡constrainedmethodiSutilizedtohandletheconstraints．Accordingtotheconstraintviolationofthecolony，thesleve1iSsettoutilizetheusefulinformationforthebetterinfeasibleindivid

4、ua1sufficientlyandtoimprovethesearchefficiencyforthefeasiblespace．Simultaneously，basedonthefeatureofeconstrainedmethod，aneworderingrulebasedon￡constrainedisusedtoobtaintheimmigrationrateandtheemigrationrate，whichcandynamicallybalancetherelationbetweenthefeasibleindividuals

5、andtheinfeasibleindividuals．Additionally，anewdynamicmigrationstrategyisshowntoenhancethesearchabilityofthemigrationmechanism．Eventually，withthepurposeofimprovingtheprecisionofconvergence，thepiecewiselogisticchaoticmapisintroducedtoimprovethevariationmechanism．Numericalexpe

6、rimentson13well—knownbenchmarktestfunctionsshowthateBB{)一dmiscompetitivewithotheroptimizationalgorithmsontheaccuracyandthespeedofconvergence，especiallywhenbeingappliedtosolvecomplexsingle—objectiveCOPs．Keywordsbiogeography—basedoptimization；econstrained；orderingrule；dynami

7、cmigrationstrategy；chaoticvariationmechanism摘要提出基于动态迁移的￡约束生物地理学优化算法(eBBOdm)．首先，利用￡约束方法来处理约束条件，并根据群体约束违反度的优劣程度对水平参数e进行自适应调整，充分利用较优不可行个体的有效信息，有效提高对可行域的搜索效率．其次，采用新的￡约束排序机制确定迁入率和迁出率，较好地平衡可行个体与不可行个体之间的关系．再次，为了增强迁移机制的搜索能力，提出新的动态迁移策略．最后，采用分段logistic混沌映射改进物种变异机制，提高了算法的收敛精度．通过对l3个标

8、准测试函数的仿真实验表明，eBB()一dm较其他算法在收敛精度和收敛速度上具有明显优势，尤其适合于复杂单目标约束优化问题的求解．关键词生物地理学优化；e约束；排序机制；动态迁移策