I am trying to get the real color of my global variable.
Here my code:
breed [players player]
globals [
INITIAL-POSSESSION ;
]
to setup
clear-all
reset-ticks
set-initial-possession
end
to go
ticks
ask players [
decision
]
end
to-report initial-possession
report random 2
end
to set-initial-possession
ifelse initial-possession = 1
[
set INITIAL-POSSESSION black]
[
set INITIAL-POSSESSION white]
end
to decision
if ([color] of INITIAL-POSSESSION) = black
[] ;;do something
if ([color] of INITIAL-POSSESSION) = white
[];;do something
end
But I get this error:
OF expected input to be a turtle agentset or link agentset or turtle
or link but got the number 0 instead.
So I change it with (and it works):
to decision
if INITIAL-POSSESSION = 0
[]
if INITIAL-POSSESSION = 9.9
[]
end
But there is any other way to do that? (I am using netlogo 6.0)
I think there may be some code missing so I can't confirm, but it looks like you may not have set BALL-OWNER up as a turtle or patch, and instead assigned a value directly to that variable. of queries a variable from an agent (or a list of variables from an agentset), so if BALL-OWNER is set to a value, NetLogo gets confused. If you do assign an agent to BALL-OWNER, however, your approach should work fine. For example, try running the code below:
to setup
ca
crt 10 [
setxy random-xcor random-ycor
set color one-of [ red blue ]
]
reset-ticks
end
to go
let ball-owner one-of turtles
ifelse [color] of ball-owner = red [
print "red team has possession"
] [
print "blue team has possession"
]
end
Edit: You can definitely use a global to pick a color just as you did in your second code block- I just wanted to point out that of is specifically tied to agents. If you want to store a color in a global variable for comparison, that's possible, it's just that your comparison is simpler than using of:
globals [ initial-possession ]
to setup
ca
crt 3
set-initial-possession
reset-ticks
end
to go
ask turtles [
decision
]
end
to set-initial-possession
set initial-possession ifelse-value ( random 2 = 1 ) [black] [white]
end
to decision
ifelse initial-possession = black [
print "I see that black has possession"
] [
print "I see that white has possession"
]
end
I'm not sure if that helps, it may depend on your purpose of storing a color in the global in the first place!
Related
I am a beginner with NetLogo and I am trying to ask some turtles (students from different social classes) to link to other turtles (schools). What I would like is for the working class students to look for the school with the highest achievement which is at the same not expensive and has not reached the max number of links allowed. If the desired school has reached the max number of links allowed I want the student to look for the next school with the highest achievement which has not reached the max number of links allowed and so on.
This is the code. I get the following message "ANY? expected input to be an agentset but got the turtle (school 1) instead."
breed [schools school]
breed [upperclass upperclass-student]
breed [workingclass workingclass-student]
upperclass-own [achievement enrolled? target]
workingclass-own [achievement enrolled? target]
schools-own [schoolachievement expensive? ]
to setup
clear-all
set-default-shape schools "house"
set-default-shape upperclass "person"
set-default-shape workingclass "person"
ask patches [ set pcolor 8 ]
create-schools num-of-schools [ setxy random-xcor random-ycor set schoolachievement random-normal 5 1
set expensive? one-of [ true false ] ]
create-upperclass num-of-upperclass [ set color white setxy random-xcor random-ycor set achievement
random-normal 5 1 ] ;Students from upper class have higher achievement
create-workingclass num-of-workingclass [ set color red setxy random-xcor random-ycor set achievement
random-normal 4 1 ]
end
to go
ask workingclass [
choose-school ]
end
to choose-school
if breed = workingclass [
set target one-of schools with-max [ schoolachievement ] with [ expensive? = false ] ]
if any? target with [ count link-neighbors < max-link-count ] [
create-link-with target ]
set enrolled? TRUE
end
Your problem is the difference between an agent and an agentset, which is a somewhat subtle problem. The with-max returns an agentset - a set of agents (in this case turtles). That agentset can have 0 members, 1 member, 2+ members but is a set even if it is empty. However, the one-of selects one agent from the agentset and returns it as an agent, not an agentset. That is, NetLogo knows anything that is returned by one-of must be exactly one agent. At this point, primitives that are for agentsets (like any?) will throw an error unless they can also be used on individual agents.
So, back to your code. I like the readability of checking whether there are viable schools and then selecting one, which is what I think you meant to do. That would be:
to choose-school
if breed = workingclass
[ set targets schools with-max [ schoolachievement ] with [ expensive? = false ]
set candidates targets with [ count link-neighbors < max-link-count ]
if any? candidates
[ create-link-with one-of candidates
set enrolled? TRUE
]
]
end
Note that I also changed to targets instead of target, which is one way to keep track of whether something is an agent or an agentset.
The other way you could do this and keep it as an agent is:
to choose-school
if breed = workingclass [
set target one-of schools with-max [ schoolachievement ] with [ expensive? = false ] ]
if target != nobody and [count link-neighbors] of target < max-link-count [
create-link-with target ]
set enrolled? TRUE
end
So you can use nobody instead of any? but you can't also use with in that line because the with is really a filter on a set.
I also think you have a bracketing issue - I assume you want set enrolled? TRUE inside the brackets. I left it in the second fix, but changed in the first error (as well as changing bracket position convention to make the code block structure more visible)
I'd need to pick an object in a bag containing 20 elements with attributes c (color) and s (size). Both color and size are numbers (e.g. c= {red = 256, black = 0, ... } = {256, 0, ...}).
As in Python I'd use random.choice in numpy library, I found on the web that the corresponding function in Netlogo is the extension rnd.
Struggling along a possible solution, I did
Edited:
breed[people person]
people-own
[
ball
size
color
bag
]
to setup
create-people 5
[ set color gray
setxy random-xcor random-ycor
]
ask people[
set bag [ ] ; 0 items
]
end
To create the balls:
to create-balls
set color random 300 ; color
set size random-float 5 ; size
let this-ball self
ask one-of people [ ; ask one of people to put the ball created into the bag
set bag fput this-ball bag ; add ball to the bag
]
end
The code below should include the part of drawing:
to draw
ask one-of people [
rnd:weighted-one-of bag [ ] ; I do not know what I'd write in the brackets
]
end
As you can easily see, I've many doubts about how to implement the code.
How can I select one item from the bag depending on its size (or color)?
Can you please help me out with it?
Here is a complete model that creates people and balls as turtle agents and has 30 of the balls get chosen weighted by their size. It then opens an inspect window for the person who has chosen the most balls.
extensions [rnd]
breed [people person]
people-own [ my-balls ]
breed [balls ball]
balls-own [ chosen? ]
to setup
clear-all
create-people 20
[ setxy random-xcor random-ycor
set my-balls (turtle-set nobody)
]
create-balls 50
[ hide-turtle
set size one-of [1 2 3 4 5]
set color one-of [red white blue yellow]
set chosen? false
]
repeat 30 [draw-ball]
inspect max-one-of people [count my-balls]
end
to draw-ball
ask one-of people
[ let bag-of-balls balls with [not chosen?]
let choice rnd:weighted-one-of bag-of-balls [size]
ask choice [set chosen? true]
set my-balls (turtle-set my-balls choice)
]
end
Some things to notice:
There are NO lists in this code. There are situations where you should use lists. Common uses include memory where the order is important (eg you only want to keep track of the last 5 other people seen) or where the same agent can appear multiple times. And the list commands are very powerful. However, unless you need a list then you should use agentsets.
Each person has their own bag called 'my-balls' that contains the balls they select. That is initialised as a turtle-list as part of the setup.
I used a variable called 'chosen?' that is owned by each ball to track whether it is still in the bag for the next person to choose. Then the bag-of-balls is created only as all the balls not yet chosen.
The code for weighted random choice (when choosing from agentsets) simply has the name of the variable holding the weight as the reporter, but you could use some function such as rnd:weighted-one-of bag-of-balls [size ^ 2] if you wanted a more complicated weighting scheme.
I'm trying to calculate and report the distance (link-length) between specific agentsets in NetLogo? Is there a way to calculate link length into a list?
The movement of the agents is based on whether the distance(connection) value is below/above a threshold. However, I'm having difficulty setting the values of the link length to variable connection. (preferably in a list). I'd appreciate any help.
globals[hourly-wage connection]
breed[offices office]
breed[employees employee]
offices-own [
pay-high ;; 7 offices pay well
pay-low ;; 3 offices dont pay well
]
to setup
clear-all
create-offices 10 [
set size 1.0
set color blue
set shape "house"
setxy random-xcor random-ycor
ask offices [create-link-with one-of other offices] ;; undirected links
ask links [set color red]
]
create-employees 2 [
set size 1
set color brown
set shape "person"
]
set hourly-wage 20
end
;;;;
to go
cal-dist
ask employees [
if connection > 15
move-to one-of high-pay office
if connection <= 15
move-to one-of low-pay office
]
end
to cal-dist
set connection [list print link-length] ;;
ask links [show link-length]
set salary (hourly-wage * connection) ;;; salary printed in a list
end
Not exactly sure what you're trying to do here with connection etc, but you can put any link variables into a list using of- for example:
to setup
ca
; First create the agents
crt 5 [
while [ any? other turtles in-radius 5 ] [
move-to one-of neighbors
]
set color blue
set shape "house"
]
; Once they're created, have them link with
; one of the other agents
ask turtles [
create-link-with one-of other turtles [
set color red
]
]
let link-lengths [ link-length ] of links
print link-lengths
reset-ticks
end
I don't know that this actually answers your question, so you may want to provide more detail as to what you're trying to accomplish with these links.
I'm modeling territory selection in NetLogo, and would like my turtles to calculate the perimeter of their territory once established. I've been trying to come up with ideas for how to do this, but haven't found a good means yet. Any ideas?
patches-own
[ owner ] ;; patches know who owns them
turtles-own
[ territory ;; agentset of patches I own
food ;; food acquired in my territory
threshold ] ;; food required, will build territory until meet this
to go
tick
ask turtles [ build-territory ]
end
to build-territory
if food > threshold [ calculate-perimeter ] ;; stop building when enough food
pick-patch ;; keep picking patches until meet threshold.
end
to calculate-perimeter
;; what could I use to add up the perimeter of the territory?
end
Thanks in advance for any suggestions!
A modification of my last answer to you:
to setup
ca
ask patches with [pxcor > 0 ] [
set pcolor white
]
crt 1
end
to go
ask turtles [
let blacklist patches with [ pcolor = black ]
let northpatches patches with [ pycor > 0 ]
let northred ( northpatches with [ member? self blacklist = false ] )
ask northred [ set pcolor red ]
let border northred with [ any? neighbors4 with [ pcolor != red ] ]
ask border [
set pcolor blue
]
print count border
]
end
You can designate border/perimeter patches as any of you territory patches with neighbors that are not territory. For you it might look something like:
ask turtles [
print count territory with [ any? neighbors4 with [owner != myself ]
]
]
Again, I can't test it without your setup so you would have to modify.
Edited below
To count the edges of patches that are on the border, you could have them count their neighbors4 that belong to another turtle. Then, they can add them to that turtle's perimeter length. For example:
to assess-perimeter ;;; must be called by a turtle
print ("Assessing perimeter")
let current-turtle who
let temp-per-len 0
let border-patches patches with [ owner = current-turtle and any? neighbors4 with [ owner != current-turtle ] ]
show (word "I have " count border-patches " border patches")
ask border-patches [
;; One way to get each border patch to check each of its neighbors
let nobodies 4 - count neighbors4 ;; if any patches are on the edge of the world, returns the number of those edges
let non-territory-edges count neighbors4 with [ owner != current-turtle ]
let border-edges nobodies + non-territory-edges
set temp-per-len temp-per-len + border-edges
]
show (word "My perimeter length: " temp-per-len )
set perimeter-length temp-per-len
end
If that is called after all turtles have chosen their entire home range, the idea is that each turtle assesses the border of its home range. Then, it has each of those border patches count its neighbors4 that have a different owner. I used "temp-per-len" as a summing variable within the loop, which is then used to set the turtles-own "perimeter-length". Full model code, including setup and definitions, here. Note- you'll have to download or copy the code, the model is too bulky to run well in the HTML format.
Also, I didn't actually count to make sure this worked perfectly- I did a quick version and crossed my fingers, but I think the idea makes sense and hopefully gets you started.
I'm new to NetLogo and am attempting to model home range selection of subsequent colonizers. The model should follow simple steps:
Individual 1 picks a home range (a subset of patches).
When individual 1 is done picking its home range, it hatches new
individual 2.
Individual 2 picks a home range, then hatches individual 3.
Individual 3 picks a home range, and so on.
I'm having trouble figuring out how to get this to work. I can get the first turtle to pick a home range. But the offspring do not. Writing the code numerous ways has only accomplished two unintended outcomes. Either endless new individuals are hatched simultaneously, before the first turtle has a home range, and the new turtles fail to pick a home range. Or, the first turtle picks its home range and hatches a new turtle, but that new turtle doesn't pick a home range. Neither outcome is what I want.
How do I set this up to run as intended, so that hatchlings pick home ranges too? Here is one simplified version of my code:
to setup-turtles
crt 1
[setxy random-xcor random-ycor]
end
to go
ask turtles [pick-homerange]
tick
end
to pick-homerange
while [food-mine < food-required] ;; if not enough food, keep picking patches for home range
[;; code to pick home range until it has enough food; this is working okay
]
[;; when enough food, stop picking home range
hatch 1 fd 20 ;; now hatch 1, move new turtle slightly away
]
end
So it is at this last part, once the home range is built, that I want a new turtle to hatch from its parent. I then want that turtle to repeat the pick-homerange procedure. How could that be coded to happen? I've tried writing this every way I can think of; nothing is working. Thanks in advance for any assistance!
One way to do this is to have each patch equal one "food value", and have turtles grow their home range until their home range supplies them with enough food. I would set this up so that patches "know" to which turtle they belong, and so that turtles know how much food they need, which patches are part of their home range, and the food supplied by their homerange. Example patch and turtle variables would then be:
patches-own [
owned_by
]
turtles-own [
food_required
my_homerange
homerange_food
]
Then, your turtles can add patches into their home range until they hit their "food_required", whatever you set that as. For simplicity, in this example I assume that turtles are territorial and so won't "share" home ranges. Further explanation of steps is commented in the code below. This is intended just to get you started- for example, it will hang if you run pick-homerange too many times.
to setup-turtles
crt 1 [
set size 1.5
setxy random-xcor random-ycor
set food_required 5 + random 5
set homerange_food 0
set my_homerange []
]
end
to pick-homerange
ask turtles [
;; Check if the current patch is owned by anyone other than myself
if ( [owned_by] of patch-here != self ) and ( [owned_by] of patch-here != nobody ) [
;; if it is owned by someone else, move to a new patch that is not owned
let target one-of patches in-radius 10 with [ owned_by = nobody ]
if target != nobody [
move-to target
]
]
;; Now add the current patch into my homerange
ask patch-here [
set owned_by myself
]
set my_homerange patches with [ owned_by = myself ]
;; calculate the number of patches currently in my homerange
set homerange_food count patches with [owned_by = myself]
;; Now grow the homerange until there are enough patches in the homerange
;; to fulfill the "food_required" variable
while [ homerange_food < food_required ] [
let expander one-of my_homerange with [ any? neighbors with [ owned_by = nobody ] ]
if expander != nobody [
ask expander [
let expand_to one-of neighbors4 with [ owned_by = nobody ]
if expand_to != nobody[
ask expand_to [
set owned_by [owned_by] of myself
]
]
]
]
;; Reassess homerange food worth
set my_homerange patches with [ owned_by = myself ]
set homerange_food count patches with [owned_by = myself]
]
ask my_homerange [
set pcolor [color] of myself - 2
]
;; Now that my homerange has been defined, I will hatch a new turtle
hatch 1 [
set color ([color] of myself + random 4 - 2)
]
]
end