基于有限理性模型的游戏智能行为系统的-.研究与实现

《基于有限理性模型的游戏智能行为系统的-.研究与实现》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、基于有限理性模型的游戏智能行为系统的研究与实现========!==!!!==!=!========================!========!!!=====!==!!!!!!=詈!!=!==!!!!!===========!!!!!!=!========!!!!!!!!!=======AbstractInthefaceofthegraduallymaturedanddiversifiedgamemarket，thegamedevelopersexpectthatthemeliorationoftheintelligencedesignwillenhan

2、cethegame’Splayability．Apartfromtheexternalperformanceofthebehaviorandtheoptimizationofthepath—findingalgorithm，itisthemainconcernofgamedevelopersthattheintelligentagentisreasonablycontrolledintheintelligentdesign．Thedevelopersnarrowthegapbetweenvirtualandrealtoimprovethegamechalleng

3、ingthroughmakingtheintelligentagentsimulateplayers’behaviorandstrategies．Thispaperchoosesthemodel—basedintelligentbehaviorsystemastheoriginalsystem，andreferencestheSOARcognitivesciencemodelandtheboundedrationalitymodeltomaketheintelligentagentfittheplayers’behaviorpatternfrom3aspects

4、：theperceptron，thedecision—makingandthebehaviorcontrolunit．Themainworkandachievementshavebeendiscussedasfollows：(1)Atfirst，thispaperfirstreconstructstheintelligentagent’Sperceptiondomainthroughintroducinginthereceptivefieldofthebiologicalvisualperceptionmodel，inordertosolwetheproblem

5、whichexitsinusingcone-horizonbooleantosettheperceptiondomain．Atthesametime，thevisualsensitivitymodelisexploitedtocalculatethevalueoftheuncertainperceptiondomain．Therefore，theinputinformationwillbedeterminedbycomprehensivelyanalyzingthepreviouslydata．Theconfrontationalexperimentsshowt

6、hatthenewperceptroncanprovideamoreadequateandusefulinformationforthedecision—makingwithoutaffectingthesystemefficiency．(2)Inthispaper，therecursivepredictionmethodcombiningtheforward—lookingtreeisproposedtoimprovethedecision—makingability．Ineverytimethesystemupdate，thismethodautomatic

7、allybuildsaforward—lookingtreebasedontheintelligentagent’Sattributes．Meanwhile，thesystemcalculatestheutilityvalueofeveryexecutorinthistreeaccordingtothepreferencerelationandtheutilityfunction，andthenselec，tsthemostsatisfactorybehaviorasoutputthroughsystematicassessingthecomprehensive

8、utilitiesoft