FIT3094-ArtificialLife,ArtificialIntelligence,andVirtualEnvironments

ASSIGNMENT2[25%]-IndividualAssessment

GoalOrientedActionPlanning

SUBMISSIONDUE:FridayWeek8,11:55PM(SubmissionA),FridayWeek10,11:55PM

(SubmissionB)

CoversTopicsfromWeeks5-7

LearningOutcomes

Completionofthisassessmentitemshoulddemonstratecompetencyinthefollowinglearning

outcomes:

●

Apply-throughpractice-basedlearning-design,development,executionandvalidationof

real-timeinteractivesoftwareusingAItechniques

●

EvaluateandapplyAIand/orALifesoftwaretechniquestomodelsimpleintelligent

behaviourindiscreteandcontinuoussimulationsandgames

●

ExplaintheconceptsandtheoriesunderlyingAIandArtificialLife;

Brief

Forthisassignment,youwillcoordinateateamofagentsusingGoalOrientedActionPlanning.Your

objectiveistocomputeplansthatachievethehighestpossiblescore.Youwillbeprovidedwith

basecodeforthisproject,andyourjobwillbetodesignasuitableplanningmodelandalgorithms.

SubmissionRequirements

Thisassignmentwillbesplitintotwoseparatesubmissions,eachfocusingonadifferentaspectof

thelearningoutcomes.Thereare100marksavailableacrossbothsubmissions.

Thefirstsubmission(SubmissionA)willbefocussedoncreatingHighLevelActions(HLActions)

andcreateaplanusingaFiniteStateMachine,worth20%oftheoverallassignmentmarksand

dueonFridayWeek8at11:55pm.

Thesecondsubmission(SubmissionB)willbefocussedontheutilisingGoalOrientedAction

Planning(GOAP)fortheseplansinstead,aswellasawrittenreport,worth80%oftheoverall

assignmentmarksanddueonFridayWeek10at11:55pm.

SubmissionsaremadeviaMoodle.ForeachsubmissionyouwillsubmityourUnrealprojectwithall

ofyourimplementationsourcecode.ForSubmissionByouwillalsosubmitawrittenreport.

EachsubmissioncomprisesasingleZIPfileandmustmeetthefollowingrequirements:

●

Thefilenameisoftheform:FIRSTNAME_LASTNAME_STUDENTID.zip

●

Afoldercalled“code”whichcontainsonlythefollowingUnrealfiles/directories:

○

Configfolder

○

Contentfolder

○

Sourcefolder

○

.uprojectfile

●

Afoldercalled“report”containingasinglePDFfile(OnlyforSubmissionB)

1

FIT3094-ArtificialLife,ArtificialIntelligence,andVirtualEnvironments

TaskOverview

Youwillhavetocoordinateanumberofdifferenttypesofagentstoperformdifferentactionsto

achieveacommongoal.

YourtaskistoimplementalgorithmsthatapplyFSMPlanningandGoalOrientedActionPlanningto

assisttheshipstoachievethelargestscorepossible.

Markswillbeallocatedfor:

1.FSMPlanning(PartA)

2.GOAPEffectiveness(PartB)

3.SystemTest(PartB)

4.WrittenReport(PartB)

Detailoneachoftherequirementsforpoints1-4aboveandthemarksallocatedtoeach

component,appearinsectionsbelow.

Youhavebeenprovidedwithbasecodethatcontainsstarterfilestoassistyouindevelopingyour

applicationofMulti-AgentPlanning.Thebasecodedoesthefollowing:

1.LoadsamapfileandconvertsitintoanUnreallevel

2.Loadsascenariofileandspawnsshipsbasedonthescenario

3.Handlesshipmovementbasedonapathgiventoit

4.IncludesGoalOrientedActionPlannerhigh-levelbase

5.IncludebaseactionforusewithGoalOrientedActionPlanner

6.Outputspointdatatohelpyouwithfine-tuningyouralgorithms

Thebasecodedoesnothandlebehaviourforavoidingcollisionsandreplanning.Thismustbe

performedaspartofyourimplementationofthecollisionalgorithms.

Inourvirtualworld,thecrewsofdifferentshipshavedifferentspecialities.Theymustworktogether

tobuildasettlement.Inthisprocess,pointsareawardedforgatheringresourcesandconstructing

buildings.

Youwillhaveatotalof180secondstotrytoachievethemostpointspossible.

TypesofAgents

Thereare4differenttypesofagents.Eachwithadistinctspecialitywhichyoucanfindinsections

below.Differentamountsofeachexistintheworld(x).Theseare:

●

Woodcutter(5)

●

Stonemason(5)

●

Farmer(5)

●

Builder(3)

2

FIT3094-ArtificialLife,ArtificialIntelligence,andVirtualEnvironments

ResourceLocations

ResourceLocationsarescatteredthroughouttheworld.EachResourceLocationcontains5unitsof

thatresourcebeforeitisdepletedanddeletedfromtheworldandshouldbereplacedbyShallow

Water.EachResourcetakesacertainamountoftimetogather.Oncegathered,Resourcesare

addeddirectlytotheAgent’sinventory,andAgentscannotperformotheractionsthandepositing

whilehavingResourcesintheirinventory.EachResourcegivesaspecifiednumberofpointswhen

depositedataHomenodeandaddsthatmanyresourcestothetotal.DepositingaResourcetakes

1second.Thereare3differenttypesofResources:

●

Wood:Takes2secondstogather,andgives1pointperunitdeposited.

●

Stone:Takes3secondstogather,andgives2pointsperunitdeposited.

●

Grain:Takes5secondstogather,andgives3pointsperunitdeposited.

GatheringResources

AllAgentscangatherresourcesfromresourcetilesonthemap.TheefficiencyofeachAgent

dependsontheirspecialisation(seebelow)whichaffectshowquicklytheycangatherresources

andhowmuchtheycancarry.

●

Woodcutter:GathersWood50%quicker,andgathers2xunitsatonce

●

Stonemason:GathersStone50%quicker,andgathers2xunitsatonce

●

Farmer:GathersGrain50%quicker,andgathers2xunitsatonce

Buildings

Todevelopathrivingcommunity,weneedbuildingstoprovideservices.Therearethreedifferent

typesofbuildings,eachwithadifferentnumberandtypeofResourcestoconstruct.Constructing

differenttypesofbuildingsgivesdifferentnumbersofpoints.Thesebuildingsare:

●

Market:Requires5Woodtobuildandgives50points

●

Theatre:Requires10Woodand5Stonetobuildandgives250points

●

University:Requires15Wood,10Stone,and5Graintobuildandgives1000points

ConstructingBuildings

Buildingscanonlybeconstructedonspecifictiles;BuildingSlots.OnlyBuilderAgentscan

constructBuildingsandcanonlydosowhentheyarenotcarryingresources.Youwillneedto

carefullydecidehowtobestutiliseyourlimitednumberofBuilderAgentstomaximisepoints.

DifferentBuildingsrequiredifferentamountsoftimetoconstruct:

●

Market:5seconds

●

Theatre:10seconds

●

University:15seconds

FSMPlanning(PartA)

SuggestedCompletionDate:EndofWeek8

YouwillberequiredtoutiliseandcreateaFiniteStateMachineinordertomakeaplanforeach

shiptobeabletomaximisepoints.

SeespecificpointmetricsinDetailedRubriconMoodle.

3

FIT3094-ArtificialLife,ArtificialIntelligence,andVirtualEnvironments

GOAPEffectiveness(PartB)

SuggestedCompletionDate:EndofWeek9

YouwillberequiredtoutilisetheexistingGoalOrientedActionPlannerinordertomakeaplanfor

eachshiptoworkcollaborativelyinordertomaximisepoints.

SeespecificpointmetricsinDetailedRubriconMoodle.

SystemTesting(PartB)

SuggestedCompletionDate:EndofWeek10

Youmustimplementamethodofreplanningagent’spathsandplanstominimisecollisions.

ReplanningisespeciallyimportantinthisAssignmentduetothelimitedevaluationtime,agents

usingoptimisedpathsareabletocarryoutmoreactionsandcollectmorepoints.

SeespecificsystemtestingmetricsinDetailedRubriconMoodle.

WrittenReport

SuggestedCompletionDate:EndofWeek10

YoumustensurethatyouincludeawrittenjustificationtoexplainyourGOAPmodel,planning

algorithmsandexecutionapproach.Belowaresomeaspectsofyourplanneryoumayliketo

discuss.Refertothemarkingrubricforguidelinesabouttheamountofdetailexpectedfordifferent

levelsofachievement.Rememberthereportisworth40%ofyourtotalmarks!

Planningmodel:

●

Worldvariables.Howyouusedtheprovidedworldvariablesandjustifications/descriptions

ofanyadditionalvariablesthatyoumayhaveadded.

●

Actions.Describeyourchosenactionsintermsoftheirpreconditionsandeffects.Include

justifications

●

Goals.Describeandjustifyyourchosengoalsforthedifferentagenttypes.Indicateand

justifytheirpriorities.

Planningapproach:

●

Expander.Discusshowyourexpanderfunctiongeneratessuccessors(forwardsor

backwards).Justifyyourapproach.

●

Search.Discusshowyourplannersearchesforsolutions.Ifyoumodifiedtheplannerin

someway(e.g.,toimproveperformance),justifyyourapproach.

Executionconsiderations:

●

Discusshowyouhandleplanfailure(i.e.,whenahigh-levelactiondoesnotproceedas

expected).Whatarethemainconsiderationsandhowdoesyourplannerreact?

●

Discusshowyouhandlenavigationfailure(i.e.,whenalow-levelpathfindingactiondoesnot

proceedasexpected).

4

FIT3094-ArtificialLife,ArtificialIntelligence,andVirtualEnvironments

AssessmentCriteria

●

WrittenReport-40marks

○

Descriptionofyourapproachtakentotheproblemathand

○

Communicationskillsdisplayedinreport

●

FSMPlanning-20marks

○

TotalPointsEarnedatEndof3Minutes

●

GOAPEffectiveness-25marks

○

TotalPointsEarnedatEndof3Minutes

●

SystemTest-15marks

○

WecomputetheSystemTestMetriccalculatedas:STM=[GOAPEffectivenessScore(see

above)]*[ProportionofSuccessfulReplans]

AdetailedmarkingrubricforSubmissionAisprovidedonMoodle.

AdetailedmarkingrubricforSubmissionBisprovidedonMoodle.

FeedbackInformation

Youwillreceiveamixofformalandinformalfeedbackforthisassessment.Demonstratorswill

provideinformalfeedbackinyourscheduledlabsessions.Formalfeedbackwillbeprovidedwithin

10businessdaysofsubmission

5

FIT3094-ArtificialLife,ArtificialIntelligence,andVirtualEnvironments

Wheretogethelp

Ifyouarestrugglingwithanythingorneedadditionalsupportpleasedonothesitatetoreachout

usinganyofthefollowingsupportservices:

Englishlanguageskills

ifyoudon’tfeelconfidentwithyourEnglish.

●

TalktoEnglishConnect:https://www.monash.edu/english-connect

Studyskills

Ifyoufeellikeyoujustdon’thaveenoughtimetodoeverythingyouneedto,maybeyoujustneeda

newapproach

●

Talktoanacademicskillsadvisor:https://www.monash.edu/learnhq/consultations

Thingsarejustreallyscaryrightnow

Everyoneneedstotalktosomeoneatsomepointintheirlife,nojudgementhere.

●

Talktoacounsellor:https://www.monash.edu/health/counselling/appointments

(friendly,approachable,confidential,free)

Thingsintheunitdon’tmakesense

Evenifyou’renotquitesurewhattoaskabout,ifyou’renotsureyouwon’tbealone,it’salways

bettertoask.

Askintheforums,book/attendaconsultationoremailyourdemonstrator

●

ForumsLink:https://lms.monash.edu/course/view.php?id=140783§ion=3

●

ConsultationSchedule:https://lms.monash.edu/course/view.php?id=140783§ion=2

●

TutorContactInformation:https://lms.monash.edu/course/view.php?id=140783§ion=0

Idon’tknowwhatIneed

EveryoneatMonashUniversityisheretohelpyou.Ifthingsaretoughnowtheywon’tmagicallyget

betterbythemselves.Evenifyoudon’texactlyknow,comeandtalkwithusandwe’llfigureitout.

Wecaneitherhelpyouourselvesoratleastpointyouintherightdirection.

6