timriffe

# the point of this script is to demonstrate the SchellingSegregation() function,
# which takes the output from SchellingProcess() and calculates the segregation index given a certain neighborhood size, by default 4x4 cells blocks in the larger matrix.
# we first source two git gists containing the functions we'll use:

# source Schelling Process functions:
# get Schelling Process functions:
source("http://raw.github.com/gist/1910183/785efd5809a05df73de23c4fa9d897e3789444ba/SchellingProcessFunctions.r")
# get Segregation Index functions for Schelling Process output:
source("http://raw.github.com/gist/1910173/248483abb658f73acb03e897d3374b5b709c3c21/SegregationIndex.r")

timriffe

# this code does a little experiment to see how the proportion living in a 500x500 game of life board changes 
# over time by starting population density.

# we first define 2 functions (from Vadim Vinichenko) for the game of life. 
# His functions are convenient because they scale well.

shiftMatrix <- function(mx, dr, dc) {
	#Shift the matrix by dr (delta r) rows and dc columns
	#by adding e.g. dr rows of zeros and removing dr rows from the other side 
	

timriffe

MontyHall <- function(verbose=TRUE){
	choices <- as.character(1:3)
	rightchoice <- sample(choices,size=1)
	yourfirstchoice <- select.list(choices=choices,multiple=FALSE,title="First Guess")
	remaining <- choices[!choices %in% yourfirstchoice]
	eliminate <- sample(choices[!choices %in% c(rightchoice,yourfirstchoice)],size=1)
	otherdoor <- choices[!choices %in% c(eliminate,yourfirstchoice)]
	if (verbose){
		cat("You chose door",yourfirstchoice,"\n")
		cat("Of the other two doors, I can tell you that it's definitely NOT door",eliminate,"\n")

timriffe

# Author: Tim Riffe tim.riffe@gmail.com
###############################################################################

# ----------------------------------
# a big fat game of life simulation 
# ----------------------------------

# --------------------------------------------------------------------------
# Explanation of what this is all about:
# --------------------------------------

timriffe

### FIXED!!

# It turns out that the assumptions you make for border have a huge impact on the evolution of a 
# Game of Life board. Our first implementations (Vadim and myself both) were to assume dead cells on the 
# borders. Here I've included a new function where edge cells have neigbors on the opposite side of the board
# i.e. the board wraps around both top to bottom and right to left, which is a shape called a taurus.

# this function runs about twice as fast as the previosu best-performer from Vadim and gives VERY different 
# results (just looking at one output, measurement, the proportion of cells alive at any given iteration).

timriffe

# FIXED!!

# 2 simulations run, both take a long time. This code will dish it out to 8 cores.
# I will post the 2 data objects produced as .Rdata files somewhere so you can just load them
# and follow the code to reproduce the figures, etc.

# rnd_threshold are starting densities:
rnd_threshold <- seq(.01,.99,by=.01)

ParThreshold <- function(thr,n_rows=500,n_cols=500,n_cycles=5){

timriffe

# this syntax includes 3 functions:

# GOLborderwrap(), which is a repeat from earlier GOL posts. Just give it a matrix and it'll iterate once
# for you

# Illustrate() uses GOLborderwrap() inside. Just give Illustrate() a matrix and it'll do 20 iterations
# by default. You can change that number using the 'iter' argument. Also you can change the speed using
# the 'frame.pause' argument.

# finally theres the GOLdraw() function, which is interactive. If you see a GOL pattern online or in a 

timriffe

# Author: Mazdaika
###############################################################################


## should be col*row, not col+row

# create matrix - initial life situation
crmtx <- function(col=5, row=5, st=0.4){
	x <- matrix(rbinom(col+row,1, st), ncol=col, nrow=row)
	return(x)

timriffe

# Author: triffe
###############################################################################

# this code will try to implement an example percolation model, where density and matrix size are parameters

# the main function also produces neat statistics along the way for post-analysis
# dead borders. either rook or queen neighborhoods work.

# um, name means percolation, with dead borders
PercIteration <- function(M, neighborhood.type = "rook"){

timriffe

# Author: Tim Riffe
# this code (almost entirely commented out because it would take so long to re-run [however, you have the 
# option!]) shows how the below results were derived. This percolation code uses a stripped-down version of 
# the previous percolation code here: https://gist.github.com/2044583 - it only produces 1 result: does the 
# infiltration reach the other side of the board or not, using the rook definition. The code was reduced to 
# be as concise as possible and then compiled, to run as fast as possible (though I'm sure there's an even 
# faster way)

# the flow: we want to try out the same percolation model for each starting density from 0 to 1 in increments
# of .001 on a 100x100 board. Each density we repeat 300 times. Since the output is TRUE/FALSE (traversed or