基于多目标优化算法的发酵过程控制方法.研究

《基于多目标优化算法的发酵过程控制方法.研究》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、一些!坠生!．．●_---___________-——————___————————-●__—————————-●-————————————●-________________________--_—_——————————————————一ResearchonOptimalControlMethodforFermentationProcessBasedMulti—objectiveOptimizationAlgorithmABSTRACTTheoptimizationoffermentationprocess

2、isofgreateffecttoimprovetheproductionefficiency．Thefermentationprocessisanonlinearandtime—varyingsystem，andtheoptimizationprobleminfermentationprocessisessentiallyamulti—objectiveoptimizationproblemwithconstraintswhichisdifficulttobesolvedbytheoptimalcontro

3、ltheoryandtraditionaloptimizationmethod．Multi—objectiveparticleswarmoptimizationalgorithmbasedonrandomsearchstrategyisverysuitabletosolvecomplexnonlinearmulti—objectiveoptimizationprobleminfermentationprocessduetoitshighsearchefficiencyandgoodrobustness．The

4、refore，itisofacademicsignificancetoresearchontheoptimalcontrolforfermentationprocessbasedonthemulti—objectiveparticleswarmoptimizationalgorithm．Basedontheanalysisofthemulti—objectiveparticleswarmoptimizationalgorithm，aself-adaptiveevolutionaryconstrainedmul

5、ti—objectiveparticleswarmoptimizationalgorithmwasproposed．GiventheinvarianceoftheParetooptimalsolutionspositionwithinanevolution，adynamiccrowdingdiversitymaintenancestrategybasedonpositioninformationwasadopted．Akindofmulti··objectivedecision·-makingmethodfo

6、rfed--batchfermentationprocesswaspresented．Onthisbasis，anon—linemulti—objectiveoptimization1II北京化1_=大学硕上学位论文controlmethodforfed-batchfermentationprocesswasgiventodeterminetheoptimalfeedprofilesperiodically．Throughtheconcreteexperimentandanalysis，thefeasibil

7、ityandvalidityofproposedmethodwasverified．Theexperimentalresultsshowthatthesearchabilityneartheconstraintsborderwithproposedself-adaptiveevolutionaryconstrainedmulti—objectiveparticleswarmoptimizationalgorithmisgreatlyincreased，andthedistributionofParetofro

8、ntisimprovedobviouslywithoutanyincreaseinthecomplexityofalgorithm．Theon—lineoptimizationcontrolmethodforfermentationprocessbasedOnmulti—objectiveparticleswarmalgorithmcanmaketheoptimizationresultbetter