NetLogo Implementation of Evacuation Scenario

Introduction to NetLogo

Intelligent Systems, Interaction and Multimedia Seminar

2012/2013

Outline



Introduction to NetLogo



Turtles, Patches, and others



GUI



Programming Concepts



Extensions & Tools



A simple example

Introduction to NetLogo (I):

What is NetLogo



A programmable modelling environment for simulating natural and social phenomena (Uri Winlensky 1999)



Agent-based M&S tool



Well suited for modelling complex systems



Hundreds or thousands of independent agents operating concurrently



Exploring the connection between the micro-level behaviour of individuals and the macro-level patterns that emerge from the interaction of many individuals

Introduction to NetLogo (I):

What is NetLogo



Easy-to-use application development environment



Quickly testing hypotheses about self-organized systems



Open simulations and play with them



Large collection of pre-written simulations in natural and social sciences that can be used and modified



Simple scripting language



User-friendly graphical interface

Introduction to NetLogo (II):

The World of NetLogo



NetLogo consists of agents living in a 2-D world divided into a grid of patches



Three different type of agents plus one more



Turtles, are the agents that move around the world



Patches

, are the pieces of “ground” on which turtles can move



Links, are agents that connect two turtles



Observer, is an agent without location that oversees everything going on in the world.



Ask agents to perform a command



Collects data from models

Patches, Turtles, System



Patches: Elements of space



Change



Do not move



Turtles:

“Social” actors



Change



Mobile



All turtles and patches put together



Typically, we wish to observe the system



How many turtles are sick? Alive?

“Rules”



Turtles and patches have rules that can



Change themselves (reflexive)



Change other turtles



Change other patches

Rules for Turtles



Reflexive behaviour

 ask turtles [ forward 1 ]



Reflexive state

 ask turtles

[ if (sick?) [ set color blue ] ]



Change other turtles



If (sick?) [ ask turtles here [ set sick? true

set color blue] ]



Change patches

 ask turtles if (sick?)

[ ask patch-here [ set grass grass

– 5 ]]

Rules for Patches



Reflexive state: patches change themselves

 ask patches [set grass grass + 1 ]



Change other patches

 ask patches in-radius 1 [ set grass 0.1 * my-grass ]



Change turtles

 ask turtles-here [ set sick? true

set color blue ]



Tself



Pself



T-to-T



P-to-P



T-to-P



P-to-T

in Summary

Introduction to NetLogo (III):

GUI - Controls, Settings, Views


GUI - Controls, Settings, Views



controls (BLUE) - allow to run and control the flow of execution

 buttons



command centre



settings (GREEN) - allow to modify parameters

 sliders

 switches

 choosers



views (BEIGE) - allow to display information

 monitors

 plots

 output text areas

 graphics window


GUI - Controls



Controls - allow to run and control the flow of execution



Buttons



Command center



Buttons - initialize, start, stop, step through the model

 “Once” buttons execute one action (one step)





Command center - ask observer, patches or turtles to execute specific commands during the execution

Introduction to NetLogo (IV):

GUI - Settings



Settings - allow to modify parameters



Sliders



Switches



Sliders

- adjust a quantity from min to max by an increment



Switches - set a Boolean variable (true/false)



Choosers - select a value from a list

Introduction to NetLogo (V):

GUI - Views



Views - allow to display information



Monitors



Plots



Graphics window



Monitors - display the current value of variables



Plots - display the history of a variable’s value

Introduction to NetLogo (V):

GUI - Views



Graphics window - T he main view of the 2-D NetLogo world

Adjust speed right-click brings up turtle/patch inspector

Introduction to NetLogo (VI):




Agents



Procedures



Variables



Ask



Agentsets



Breeds



Synchronization


Programming Concepts - Agents



Each agent can carry out its own activity, all simultaneously



Patches



Form the 2D world

– They don’t move, but they sense



They have integer coordinates (pxcor, pycor)



Can generate turtles



Turtles

 move on top of the patches

 have decimal coordinates (xcor, ycor) and orientation (heading)



Observer



Can create new turtles



Can have read/write access to all the agents and variables


Programming Concepts - Procedures



Procedures tell agents what to do



Command is an action for an agent to carry out



Usually begin with verbs

to setup

clear all

create 10

end to draw-polygon [ num-sides size ]

pd repeat num-sides

[ fd size rt (360 / num-sides) ]

end


Programming Concepts - Procedures



Reporter computes a result and report it



Usually begin with nouns or nouns-phrases

to-report absolute-value [ number ]

ifelse number >= 0

[ report number ]

[ report 0 - number ]

end



Procedures: Commands or Reporters implemented by the user



Primitives: Commands or Reporters built into

NetLogo(language keywords)



– Variables (i)



Variables



Global variables



Turtle & patch variables



Local variable



Global variables



Every agent can access it



Only one value for the variable



Turtle & Patch variables



Each turtle/patch has its own value for every turtle/patch variable



Local variables



Defined and accessible only inside a procedure



Created by the command let



– Variables (ii)



Built-in:



Turtle variables: color, xcor, ycor, heading, etc



Patch variables: pcolor, pxcor, pycor, etc



Defining global variables:



global [ clock ]



Defining turtle/patch variables:



turtles-own [ energy speed ]



patches-own [ friction ]



Defining a local variable:

 to swap-colors [ turtle1 turtle2 ]

let temp color-of turtle1

….


Programming Concepts - Ask



Ask - specify commands to be run by turtles or patches



Examples

 asking all turtles:



ask turtles [ fd 50 ... ]

 asking one turtle:



ask turtle 5 [ ... ]

 asking all patches



ask patches [ diffuse ... ]



Only the observer can ask all turtles or all patches



– Agentsets (i)



Agentset - definition of a subset of agents



Contain either turtles or patches



Is in a random order



Allows to construct agentsts that contain some turtles or patches



Example:

 all red turtles:



turtles with [ color = red ]

 all red turtles on the patch of the current caller (turtle or patch):



turtles-here with [ color = red ]

 all patches on right side of screen:



patches with [ pxcor > 0 ]

 all turtles less than 3 patches away from caller (turtle or patch):



turtles in-radius 3



– Agentsets (ii)



Using agentsets

 ask such agents to execute a command



ask <agentset> [ ... ]

 check if there are such agents



show any? <agentset>

 count such agents



show count <agentset>

 example: remove the richest turtle (with the maximum

“assets” value)



ask max-one-of turtles [ sum assets ] [ die ]


Programming Concepts - Breeds



Breed - a “natural” kind of agentset



Different breeds can behave differently

 breed [wolves wolf]

 breed [sheep a-sheep]



A new breed comes with automatically derived primitives:



create-<breed>, create-custom-<breed>, <breed>-here, <breed>-at



Breed is a turtle variable



ask turtle 5 [ if breed = sheep ... ]



A turtle agent can change breed



ask turtle 5 [ set breed sheep ]


Programming Concepts - Synchronization



Agents run in parallel (each agent is an independent thread)

 asynchronous commands:



ask turtles [ fd random 10

do-something]

 Agent threads wait and “join” at the end of a block

 synchronous commands:



ask turtles [ fd random 10 ]



ask turtles [ do-something ]

René Doursat, 2008

René Doursat, 2008

Introduction to NetLogo (VII):

Extensions & Tools



Extensions Guide



Sound



Robotics/NetLogoLab



GIS



Bitmap



Quicktime for Java



BDI architecture FIPA



Applets



Shapes Editor



Behaviour Space



System Dynamics



HubNet



Logging



Controlling



Mathematica link



NetLogo 3D

NetLogo References



NetLogo user manual http://ccl.northwestern.edu/netlogo/docs/



Agent-based and Individual-based Modeling: A Practical Introduction, by

Steven F. Railsback and Volker Grimm (NetLogo v5.0)



NetLogo Learning Lab http://www.professorgizzi.org/modelingcomplexity/netlogo/index.html



NetLogo 5.0

– Quick Guide, Luis R. Izquierdo



Fundamentals of Multi-agent Systems with NetLogo Examples, José M.

Vidal http://multiagent.com/p/fundamentals-of-multiagent-systems.html



Origins of Life: From Geochemistry to the Genetic Code http://origins.santafe.edu/tutorials/netlogo

A simple tutorial



Create via “File/New”, a new NetLogo program



Save it, via “File/Save as” with the name

MushroomHunt.nlogo



From the “Settings” button

 view of the World’s geometry



To initialize the World and run the model

 setup procedure

 go procedure

1



“Interface” tab -> “Button”

 create setup button

 similarly create a go button

2



In “Code” tab



Create the skeleton of setup & go



Change setup to



Create the clusters of mushrooms (patches).



The cluster can be a model parameter



Define a global variable num-clusters

 Modify the setup to turn in red randomly a “num-cluster” patches

 create the turtles

 use the primitive create-turtles

3

4



In the go procedure



Tell to turtles what to do. In this case to search for mushrooms



So we need a search procedure



Let’s define search.



After globals statement define

5



We update the setup procedure

6

 and the search procedure as well as

7

The modelling cycle for the Mushroom-hunter problem

1.

Formulate the problem



What search strategy maximizes the rate of finding items if are distributed in clusters?

2.

Formulate hypothesis for essential processes and structures

 process switches from large-scale movements to small-scale searching depending on previous

 discoveries

3.

Choose scales, entities, state variables, processes and parameters

4.

Implement the model

5.

Analyse, test and revise the model

 we could the model by trying different search algorithms and parameter values analyse to see which produces the highest rates

No results