I'm trying to set up a toy model of connected turtles. Every turtle has a random belief [0,1] and I want to implement some conditions:
If it's a turtle with belief > .5 the turtle doesn't do anything.
If it's a turtle with belief < .5 and one of their immediate neighbour has a belief > .5, their new belief (of the turtle with belief < .5) would be belief + .1.
If the turtle with belief < .5 doesn't have an immediate neighbour with belief > .5, the turtle doesn't do anything.
The basic setup:
turtles-own [belief]
to setup
ca
reset-ticks
crt 5 [
set shape "person"
layout-circle turtles 5
set belief random-float 1
]
ask turtles [ create-links-with min-n-of 2 other turtles [distance myself] ]
end
This creates a 5 turtles cycle network. Now they have to check the beliefs of their (two) closest neighbors. I tried this:
to go
ask turtles with [one-of link-neighbors with [belief > .5]] [
set belief belief + .1
]
end
So the basic idea is that every turtle with a connected neighbour with belief > .5 has to add .1 to its own belief. I have tried several variations of the ask turtles with line, but it always shows me a runtime error of this sort: WITH expected a true/false value from (turtle 0), but got (turtle 2) instead. Any ideas? Thank you in advance.
one-of returns an agent. It is not suitable in this case for evaluating a condition with with- the condition is incomplete, since the line ask turtles with [one-of link-neighbors with [belief > .5]] translates (roughly) to "ask turtles with the condition turtle x to do something".
I think what you're after is something that uses count to count the number of believing turtles or any? to evaluate if an agentset exists:
to go-1
ask turtles with [ (count link-neighbors with [belief > 0.5]) > 0 ] [
set belief belief + 0.1
]
end
to go-2
ask turtles with [ any? link-neighbors with [ belief > 0.5 ] ] [
set belief belief + 0.1
]
end
Related
I would like a turtle to go to the closest patches with most turtles if a threshold of a given variable is met for 5 ticks.
My code is:
to move
let count-tick 5
if var >= 9.5 [
set count-tick count-tick - 1
if count-tick = 0 [
ask turtle [
let nearest-group min-one-of (patches with [sum turtles >= 3] in-radius 3 ) [ distance myself ]
move-to nearest-group ;; go to the biggest crowd near you
ask turtle [ ;; once there do the following
set shape "star"
set color red
]
]
]
]
end
The issue I have is that a) I am unsure how to say the patch with >= 3 turtles closest to you at the given range of 3 (attempted code above) and b) how to say once there, change your shape.
Revised to keep a permanent variable to track whether the variable is high enough 5 times in a row.
turtles-own
[ count-tick
]
; wherever you create the turtles, you need to `set count-tick 5`
to move
ifelse var >= 9.5
[ set count-tick count-tick - 1 ]
[ set count-tick 5 ]
if count-tick = 0
[ let nearest-group min-one-of (patches with [count turtles >= 3] in-radius 3 ) [ distance myself ]
move-to nearest-group ;; go to the biggest crowd near you
set shape "star"
set color red
]
end
First, you are already within an ask turtles code block from the procedure calling this move procedure. So you don't need the additional ask turtles. Look up ask in the NetLogo Dictionary, it iterates through the turtles, running all the code for each turtle in turn.
Second, you need count turtles rather than sum turtles as sum is to add up values.
Note that there is no error checking in this, you may have problems if there are no patches within radius of 3 that have at least 3 turtles.
Maybe i worded the question wrong but what i want to do in the code is when another turtle meets another turtle they create links with each other
I know its
to go
tick
make-circle
move
if ticks >= timer1 [stop]
end
to move
ask turtles
[
create-links-with other turtles ;here is where i want to put the code
set heading random 360 fd 1]
create-network
end
to create-network
ask links [
set thickness 0.01 * counter
if [patch-here] of end1 = [patch-here] of end2
[set counter (counter + 1)]
]
end
but im not sure how to word it correctly to link when they meet how do i do that
Establish a breed variable for your counts. Then create a link between all turtles that are on the same patch with other turtles-here. Then increment that count variable for when the others turtles have met the original calling turtle. I'll note that I increment the count variable by .5 because each turtle in the link will increment it (there's 2 turtles, so .5 * 2 = 1).
links-own [meets]
to setup
clear-all
crt 100 [setxy random-xcor random-ycor ]
end
to go
ask turtles [fd 1]
ask turtles [rt random 90]
ask turtles [ create-links-with other turtles-here]
ask turtles [ ask other turtles-here [ask link-with myself [ set meets meets + .5]]]
end
I am writing a procedure (Pass-Away-Space) to calculate mortality of turtles moving from the origin (start-patch) through out the world. Each turtle calculates its own mortality based on its distance from the origin (start-patch). The code I am attempting to implement for this procedure is as follows:
to Pass-Away-Space
ask turtles [
let chances 1 - exp( -1 * mortality * [distance start-patch] of turtles)
if chances >= 1 [die
set dead-count dead-count + 1
]
]
end
The error I am getting is expected input to be a number but got the list. I am not sure what the issue is and I was wondering if anyone could point out and rectify the problem with the code.
The problem here is your of turtles. Since an ask procedure affects one turtle at a time, each turtle in your procedure above is evaluating the [distance start-patch] of all turtles instead of just its own distance to the start patch. To clarify, check out the following setup:
globals [ start-patch ]
to setup
ca
reset-ticks
crt 10 [
setxy random 30 - 15 random 30 - 15
]
set start-patch patch 0 0
end
to incorrect-example
ask turtles [
print ([ distance start-patch ] of turtles)
]
end
to correct-example
ask turtles [
print distance start-patch
]
end
Compare the print output of the incorrect-example and the correct-example procedures, and you'll see that when you use [distance start-patch] of turtles you get the list of distances of all turtles. When you ask turtles to evaluate a turtles-own variable (including color, size, etc) each turtle will automatically access its own version of that variable- there's no need to specify which turtle. So, your pass-away-space might look something more like below (untested):
to Pass-Away-Space
ask turtles [
let chances 1 - exp( -1 * mortality * (distance start-patch) )
if chances >= 1 [
die
set dead-count dead-count + 1
]
]
end
In my model I have some agents which act as food items with a set energy. These are fed upon by a number of turtle breeds who each have their own food-energy which is less than the energy of the food item.
The code for the feeding agents is as follows:
to eat
ifelse [food-energy] of myfood > 1.5 [
set food-energy 1.5]
end
and the associated code for the food item to decay is:
to decay
if any? turtles-here [set food-energy
(1.5 * count feeders-here with [myfood = myself]
end
The problem occurs if the energy of the food is not an exact multiple of the amount of energy the feeders can consume. So for example it can go down to 1 and this results in the feeders taking 1.5 units which should be impossible. This is exacerbated when I have different breeds with different food energies (i.e. < or > 1.5).
So my question is how can I get this things to balance?
You need to study the Wolf-Sheep Predation model. This is the first NetLogo tutorial: http://ccl.northwestern.edu/netlogo/docs/tutorial1.html There are five versions of increasing complexity in the NetLogo Models Library, which are covered in chapter 4 of Wilensky and Rand (2015), which you should read.
See some related material here:
http://jasss.soc.surrey.ac.uk/14/2/5.html
Some hints follow, but many details need filling in.
breed [feeders feeder]
patches-own [ food-energy ]
feeders-own [ myfood ]
to setup
ca
ask patches [set food-energy random 50]
create-feeders 500 [
move-to one-of patches
set myfood one-of patches
]
end
to go
ask feeders [move]
ask feeders [feed]
ask patches [growback]
end
to move ;how shd they move?
rt random 20
left random 40
fd 1
;shd movement cost energy?
end
to feed
if (patch-here = myfood) [
let _extracted min (list food-energy 1.5)
set food-energy (food-energy - _extracted)
]
end
to growback
;do you want growback?
end
thanks for your responses. I'll try to implement them. This is one inelegant workaround that worked for me:
to eat
ifelse (food-energy / capacity) < 1 and [meat] of myfood > capacity [
set food-energy 1.5] [set food-energy [meat] of myfood
ask myfood [set shape "square"]]
if (food-energy / capacity) = 1 [
set color white]
if (food-energy > 0 and food-energy / capacity < 1)
[ set color white ]
end
This was initially causing a problem such that when the food energy went down to 0 and I asked it to die any animal looking at the [meat] of myfood lost it's target and I got an error. So I made the animals break this connection once their colour was white.
to ignore
set myfood nobody
set food-energy food-energy * 1
end
to setup
ca
reset-ticks
ask patches [
set inside? (abs pycor < 10 and abs pxcor < 10)
set exit? false
ask patch 11 0 [ set pcolor lime set exit? true]
]
repeat initial-population [ ; start condition turtles with any other turtles on neighbors
ask one-of patches with [
inside? and (not any? other turtles-here) and (not any? turtles-on neighbors)] [
sprout 1 [
set color blue
set size 1
]]]
end
to go
tick
define-neighbors-radius-2
move
end
to define-neighbors-radius-2
ask turtles [
set neighbors-ahead2 patches at-points [[2 1] [2 0] [2 -1]]
set neighbors-for-y-up2 patches at-points [[2 0] [2 -1] [1 -2] [0 -2] [-1 -2]] with [inside?]
set neighbors-for-y-down2 patches at-points [[-1 2] [0 2] [1 2] [2 1] [2 0]] with [inside?]
]
end
to move
;; my intent to move turtles to exit without their neighbors are occupied by other turtles, ;;that is the 8 patches around turtles are empty until exit?
ask turtles[
ifelse inside? [
if ycor = 0 [ ;strategy to turtles with in front exit
ifelse exit? [
set heading 90
fd .5
]
[
facexy 11 0
if (not any? turtles-on neighbors) and (not any? turtles-on neighbors-ahead2) [
fd .5
]
]
]
if ycor > 0 [ ; strategy to turtles occupied "bottom-side" of inside?
facexy 11 0
if (not any? turtles-on neighbors) and (not any? turtles-on neighbors-for-y-up2) [
fd .5
]
]
if ycor < 0 [ ; strategy to turtles occupied "down-side" of inside?
facexy 11 0
if (not any? turtles-on neighbors) and (not any? turtles-on neighbors-for-y-down2) [
fd .5
]
]
]
[
set heading 90
fd .5
]
]
end
I try to move turtles to exit but not all turtles move, why?
also, turtles must go out with ycor = 0, that is obliques direction don't allow because neighbors will occupied patches aren't inside!
Can't public this question because "looks like my post is mostly code", so talk about of my life:
seriously my intent is to create a crowd in front of exit and set some rules to delay the turtles flow exit, for this I need the neighbors empty to show interaction between agents.
(also accept some suggest to set this interaction) but at the moment turtles reach exit!
thanks
One problem with this code is that have you three ifs where it appears you only want one of them to run each time, but it's possible for more than one of them to trigger. You have:
if ycor = 0 [ ...commands #1... ]
if ycor > 0 [ ...commands #2... ]
if ycor < 0 [ ...commands #3... ]
but the turtle's ycor might be changed by commands #1 or #2. So it's possible that both #1 and #2 might run, or both #2 and #3, and perhaps other combinations as well. I assume you intended that each turtle should only move once per tick, so I recommend you rewrite this as:
ifelse ycor = 0
[ ...commands #1... ]
[ ifelse ycor > 0
[ ...commands #2... ]
[ ...commands #3... ] ]
Something else that should be fixed in this code:
reset-ticks should go at the end of setup, not near the start. It tells NetLogo setup is finished.
tick should go at the end of go, not near the start. It tells NetLogo a tick has finished.
I don't know if either of these problems I've pointed out are actually causing the unintended behavior you're seeing. Perhaps the bug isn't obvious just from reading over the code. In that case, you have two possible courses of action ahead of you:
1) Back up. Throw away this broken code and go back to the last version of your model that contained only code that have you have verified to work correctly. Then try again, but this time, don't add so much new code all at once. Attempt to make a very small improvement to the code, and get that working, before moving on to the next small improvement. And so on. If at any point you get stuck, come here, show your code, and ask a specific question about it. That question should be much easier to answer than your current question. Questions of the form “Here is a big mass of code that doesn't work, help!" are very difficult to answer.
2) Press on by investigating your questions yourself. You write, “not all turtles move, why?” Perhaps someone can answer this just by reading your code; I can't (unless my first guess above is correct). In order to figure it out, I would have to run the code myself and do experiments with it. I'd do things like try running it with just a single turtle and see if that case fails; add print statements to the code, showing each the turtle's coordinates and motions, so I could try and figure out which turtle is going wrong, and exactly under what conditions; and so forth. It's like detective work, or like doing chemistry experiments in the lab.