Sample Quantiles of Time-Persistent Statistics (TPS)
Computes the sample quantiles of a time-persistent statistic corresponding to the given probabilities.
quantileTPS(times = NULL, numbers = NULL, probs = c(0, 0.25, 0.5, 0.75, 1))
times |
A numeric vector of non-decreasing time observations |
numbers |
A numeric vector containing the values of the time-persistent statistic between the time observation |
probs |
A numeric vector of probabilities with values in [0,1] |
The lengths of times and numbers either must be
the same, or times may have one more entry than numbers
(interval endpoints vs. interval counts). The sample quantiles are calculated
by determining the length of time spent in each state, sorting these times,
then calculating the quantiles associated with the values in the prob
vector in the same fashion as one would calculate quantiles associated with
a univariate discrete probability distribution.
Computes the sample quantiles of the time-persistent statistic provided.
Barry Lawson (blawson@richmond.edu),
Larry Leemis (leemis@math.wm.edu),
Vadim Kudlay (vadim.kudlay@richmond.edu)
times <- c(1,2,3,4,5)
counts <- c(1,2,1,1,2)
meanTPS(times, counts)
sdTPS(times, counts)
quantileTPS(times, counts)
output <- ssq(seed = 54321, maxTime = 1000, saveNumInSystem = TRUE)
utilization <- meanTPS(output$numInSystemT, output$numInSystemN)
sdServerStatus <- sdTPS(output$numInSystemT, output$numInSystemN)
quantileServerStatus <- quantileTPS(output$numInSystemT, output$numInSystemN)
# compute and graphically display quantiles of number in system vs time
output <- ssq(maxArrivals = 60, seed = 54321, saveAllStats = TRUE)
quantileSys <- quantileTPS(output$numInSystemT, output$numInSystemN)
plot(output$numInSystemT, output$numInSystemN, type = "s", bty = "l",
las = 1, xlab = "time", ylab = "number in system")
labels <- c("0%", "25%", "50%", "75%", "100%")
mtext(text = labels, side = 4, at = quantileSys, las = 1, col = "red")
abline(h = quantileSys, lty = "dashed", col = "red", lwd = 2)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.