Reas智能體模擬軟件基礎(chǔ)資料

RepastThe

Repast

3

family

of

tools

is

the

precursor

to

Repast

SimphonyRepast

forJavaRepast

for

the

Microsoft

.NET

framework,

andRepast

for

Python

ScriptingRepast

Simphony

(Repast

S)

is

in

pure

JavaExtends

the

Repast

portfolio

by

offering

a

new

approachtosimulation

development

and

executionA

set

of

advanced

computing

technologies

for

applications

suchas

social

simulationJava

APIs

as

well

as

GUIs

for

running

and

developing

modelsand

EclipsepluginRepast

home

page:

/Several

links

on

the

right

of

the

pagePapersDocumentationOn-line

introductory

tutorial:

shows

you

how

to

use

theinterface

to

create

modelsComes

with

the

download,

installed

as…\RepastSimphony-1.2.0\docs\Getting

Started.pdfThe

HTML

version

is

in

thefolder…/RepastSimphony-1.2.0/docs/tutorial/SIM/Repast

Simphony

Development

Environment:

How

to

set

upan

Eclipse

development

environment

for

Repast.This

comes

with

the

download,

and

is

installed

as…\RepastSimphony-1.2.0\eclipse\and

the

file…\RepastSimphony-1.2.0\eclipse\readme\readme.htmlRepast

Simphony

Reference:

Coding

the

APIsThis

comes

with

the

download,

and

is

installed

as…\RepastSimphony-1.2.0\docs\Reference.pdfThe

HTML

version

is

in

thefolder…\RepastSimphony-1.2.0\docs\reference\SIMFeatures

of Repast

SimphonyFluid

model

component

development

using

any

mixture

ofJava,Groovy,

and

a

flowchart

interfaceGroovy:

a

dynamic/scripting

language

for

the

Java

Platform

as

analternative

to

JavaFeatures

similar

to

those

of

Python,

Ruby,PerlLibraries

forgenetic

algorithmsneural

networksregressionEstimate

parameters

to

give

a

best

fit

for

the

data—e.g.,

leastsquaresrandom

number

generationspecialized

mathematicsAn

automated

Monte

Carlo

simulation

frameworkRelies

on

repeated

random

saplingUsed

when

it’s

infeasible/impossible

to

compute

an

exact

resultwith

a

deterministic

algorithmAutomated

connections

to

optional

external

tools,

including

plugins

tothe

R

statistics

environment

(alternative

to

commercial

SPSS)*ORA

risk

assessment

toolGiven

social,

knowledge

and

task

network

info,

find

those

people,skills

or

knowledge

and

tasks

critical

for

performance

or

info

securityPajek

network

analysis

programJoSQL

for

performing

SQL

queries

on

collections

of

Java

objectsthe

VisAD

(Visualization

for

Algorithm

Development)

scientificvisualization

package,A

Java

component

library

for

interactive

and

collaborativevisualization

and

analysis

of

numerical

datathe

Weka

data

mining

platform,Machine

learning

algorithms

for

data

mining

tasks

that

can

be

applieddirectly

to

a

dataset

or

called

from

Java

codemany

popular

spreadsheetsMATLABFull

integration

withthe

JUNG

(Java

Universal

Network/Graph

Framework)

networkmodeling

libraryAn

extendible

language

for

the

modeling,

analysis,andvisualization

of

data

that

can

be

represented

as

a

graph

ornetworkMicrosoft

ExcelspreadsheetsUCINET

DL

file

importingA

comprehensive

package

for

the

analysis

of

social

network

data($40

forstudents)Geographical

spaces

including

2D

and

3D

GIS

supportInfo from

the

Repast

3

Website/repast_3/index.html

PapersS.

Deschamp,

Univ.

of

WellingtonRepast

to

simulate

a

knowledge

network

involving

heterogeneousagents—e.g.,

small

and

medium

enterprises,

universities,

publicresearch

institutesHow

co-operative

research

strategies

can

be

implementedamong

agents

with

different

objectives

and

rewardsHow

this

affects

the

dynamics

of

innovationHow

agents

evolve

and

convergeThe

impact

at

the

system

levelOwen

Densmore

of

Sun

MicrosystemsRepast

to

model

self

organizing

networks

within

the

context

of

theLocal

Knowledge

Networking

projectSeeks

to

establish

computing

sessions

without

central

control,thus

more

adaptive

during

failure

and

are

more

robustPropagate

requests

via

alternative

techniques

(e.g.,

rings

ofneighbors)TutorialsHowto

DocumentationThese

mostly

relate

to

using

the

APIs

and

are

all

briefThird

Party

TutorialsOf

variable

qualitySeveral

of

the

links

aredeadSee

especially

the

link

to

Tesfatsion’s

tutorial

(but

RepastJ)ExamplesJava

Web

Start

(.jnlp)

filesClick

thelinkClick

OK

in

the

resulting

window,

asking

whether

to

open

it

with

theJava

Web

Start

LauncherGet

a

control

bar

and

awindowWindow

with

tabs

forParameters:

set

the

values

for

the

model

parameters(seeexamples,

below)Custom

Actions:

controls

for

the

simulationRepast

Actions:built-inRepastfunctionality—e.g.,makingmovies,

editchartsControlbarThe

control

icons,

left

to

right:Load

ModelMultiple

Runs

StartStartStepInitializeStopPauseSetup

ModelView

ParametersExitFirst

step

a

couple

of

times,

then

start,

and

pause

to

inspect

a

givenstate,

then

step

or

start

againBasic

Grid

ExamplesHeatbugsSeethe

Swarm

description:

/examples-heatbugs.htmlParameter

tabs

selected Simulation

after

a

whileBasic

Grid

Examples

(cont.)Heatbugs

on

Hexagons

(Same

reference

as

previous)Hypercycles

(See

the

reference)Game

of

Life

(See

Wikipedia,

“Conway’s

Game

of

Life”)MousetrapSee

the

Swarm

description: /examples-mousetrap.htmlSugarScape

(The

reference

is

given,

but

also

Googleit.)Network

ExamplesJiggle

Toy

(Explanation

given)Friendship

Network

(Link

to

reference)Geographic

Information

Systems

(GIS)

ExamplesGIS

Agents

(Blank

screen)System

Dynamics

ExamplesRabbit

Population

(Brief

explanation

and

link

to

reference)Installing Repast

SimphonyGo

to/Under

heading

“The

Latest

Repast

Simphony

Release”

at

the

topofthe

page,click

on

“Windows

Installer”In

the

short

wizard-based

installation

process,

accept

all

the

defaultsRunningaRepa

stSimphonyEx

ampleAll

programs

Re

pa

stSi

mphony

Repa

stSi

mphonyIDEAcceptt

hi

sde

faultInthe

Packa

ge

Explo

rer

panel(l

eftsideofwindow),

expandtherepast.s

imphony.demosfolderTheMo

del:

simplevariation

ofpredato

r-preyb

eh

av

iorusing3

agen

tty

pe

s:wolves,s

heep

,andgras

sBotht

hew

olvesand

sheep

move

randomly

onagridMove

menth

asa

c幻玉ost:losten

ergy.Wo

lvesandshe

epnee

dtoeatfood

tor

ep

le

nishtheir

ene

rgyThey

dieoncethei

renergylevelr

eaches

0Wolv

espreyon

she

epMa

yeatth

emifth

e

2arelocatedin

the

sam

espat

ialpositi

onSheepma

yeatgrassifl

ocatedo

napatchwithlivinggrassOnceeaten,gr

assneedsto

r

e-growbefore

beingeaten

ag

ainRe-g

ro

wrate:co

untingdow

nafterthegrassha

sbeeneatenRepr

od

uctionismodeled

by

ar

ando

mpr

oces

sCr

eates

achild

fromt

heparen

tDividestheenergyofthe

pare

nta

ge

ntin

halfAssignsth

eenergye

qu

ally

tothepare

ntandc

hi

ldRu

n

RunCo

nfiguratio

nInthepanelo

nleft,clickonPr

edat

or

Pr

eyDo

n’tmo

difythedefau

ltsCl

ickRu

nSeethe4tabsbelowthepan

el

onthe

leftTheScenar

ioTre

eiscurrent

lys

electedandisdisplayedin

thepane

lClickth

eRunOptionstabMoves

abov

etheSc

en

ario

TreetabTheRunOptionspa

nelnowoccupiestheleft

sideo

fthewindowThe

Sche

duleT

ickDelayslid

er

initiallyind

ic

at

es0msec.Moveitto,say,6m

sec.No

wclic

kthePa

ra

meterstabParame

te

rsP

aneltoap

pe

arsFornow,

justacceptinitialval

ue

sCa

nch

angesuchparametersastheinitia

lnumbero

fsheep,

w

hata

sheepga

insfrom

agivenamountoffoo

d,andthe

reproductionra

teofthewolv

esBackintheScenarioTree,cli

ckADisplayTheGeneralinterf

ace

isi

ni

ti

al

lys

electedfr

omamo

ngt

he4opt

ionsavailable

inthe

leftp

an

elNon

eed

toc

hangeanyoft

hese

v

aluesTheAgent

Style

an

d

GridStyle

inter

faces

letyousp

ecifyth

e

orde

rinwhichthea

gent

iconsareo

verlai

d

inthed

isplay(an

dsomemore

adva

ncedthings)mo

difythe

s

izeof

t

he

gri

dun

itsspec

ifywhether

gridlinesaret

obed

isplayedNoneed

tom

od

ifyanyofthis.The

Schedule

…interface

l

estyoumo

difyhowt

he

displ

ayisupda

tedre

la

tivetothesimulation

.Ag

ain,

just

acceptdef

au

ltsNo

wcl

ickOKInthemodelwin

do

w,c

lickthe

StepRunic

ona

tthetop

(d

oublearrowpointingr

ight

)Thescen

arioi

sdisplayedinthe

l argep

anel

.Clickthe

StepR

unicon

several

timesto

stepthroug

hthesi

mulation

.No

wst

opthesimul

at

ionbyclick

ingtheSt

opSim

ulationicon,asquarewit

hinaci

rcle,just

righto

ftheStepRuniconInthe

Scena

rio

Treeagain,do

ublecli

ckAChartInitiall

y,t

heC

hartDataMa

ppingsint

erfaceisshow

n.GettheVisu

alSetting

interfa

cebyclic

ki

ngthe

optio

ninthel

eftpane

lNoneed

tochangean

yse

tt

in