Synchronous vs. asynchronous agent activation example

February 24, 2012
By

(This article was first published on R snippets, and kindly contributed to R-bloggers)

This time I have implemented NetLogo Voting model to verify how agent activation scheme influences the results.
The code executing the simulation is given below. It simulates two types of voter preferences encoded as 1 and -1. In this way average preference equal to 0 indicates 50/50 split. Voters are arranged on square grid with vertical and horizontal wrapping.

nei8 <- function(x, y, size) {
   base.x <- c(-1, -1, -10, 01, 1, 1)
   base.y <- c(-101, -1, 1, -1, 0, 1)
   1 + ((cbind(x + base.x, y + base.y) - 1) %% size)
}

step.syn <- function(space) {
    size <- nrow(space)
    new.space <- space
    for (x in 1:size) {
        for (y in 1:size) {
            nei.pref <- sum(space[nei8(x, y, size)])
            if (nei.pref > 0) { new.space[x, y] <- 1 }
            if (nei.pref < 0) { new.space[x, y] <- -1 }
        }
    }
    return(new.space)
}

step.asyn <- function(space) {
    size <- nrow(space)
    old.space <- space
    all.x <- rep(1:size, size)
    all.y <- rep(1:size, each = size)
    dec.seq <- sample.int(length(all.x))

    for (i in dec.seq) {
        x <- all.x[i]
        y <- all.y[i]
        nei.pref <- sum(space[nei8(x, y, size)])
        if (nei.pref > 0) { space[x, y] <- 1 }
        if (nei.pref < 0) { space[x, y] <- -1 }
    }
    return(space)
}

simulate <- function (size, is.syn, do.plot) {
    x <- rep(1:size, size)
    y <- rep(1:size, each = size)

    space <- 2 * matrix(rbinom(size ^ 2, 1, 0.5), nrow = size) - 1
    rep <- 0
    while (TRUE) {
        rep <- rep + 1
        old.space <- space
        if (is.syn) {
            space <- step.syn(space)
        } else {
            space <- step.asyn(space)
        }
        if (do.plot) {
            par(pin=c(3,3))
            plot(x , y, axes = FALSE, xlab="", ylab="",
                 col = space + 2, pch = 15, cex = 40 / size)
        }
        if (all(old.space == space)) {
            return(c(rep, mean(space)))
        }
        if (rep > 100) {
            return(c(NA, mean(space)))
        }
    }
}

It uses either step.syn function which assumes that all voters make decision at the same time in each time step and step.asyn where all voters are activated in random order.

To reproduce graph similar to NetLogo original the do.plot option should be set to TRUE. For example simulate(32, FALSE, TRUE) generates the following picture:


In order to compare synchronous and asynchronous voter activation regimes I have run the following code:

s.r <- replicate (1024, simulate(32, TRUE, FALSE))
as.r <- replicate (1024, simulate(32, FALSE, FALSE))

par(mfrow = c(1, 2))
boxplot(cbind("synchronous"=s.r[1,],"asynchronous"=as.r[1,]),
        main = "time")
boxplot(cbind("synchronous"=s.r[2,],"asynchronous"=as.r[2,]),
        main = "mean preference")

As shown on the picture below the distribution of mean voter preferences is very similar. However convergence speed of both methods varies greatly. Moreover under synchronous activation around 0.5% of simulations do not reach stable state.

To leave a comment for the author, please follow the link and comment on his blog: R snippets.

R-bloggers.com offers daily e-mail updates about R news and tutorials on topics such as: visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...



If you got this far, why not subscribe for updates from the site? Choose your flavor: e-mail, twitter, RSS, or facebook...

Comments are closed.