round_stochastic: Round a stochastic vector or a row-stochastic matrix

In pomdp: Infrastructure for Partially Observable Markov Decision Processes (POMDP)

Round a stochastic vector or a row-stochastic matrix

Description

Rounds a vector such that the sum of 1 is preserved. Rounds a matrix such that the rows still sum up to 1.

Usage

round_stochastic(x, digits = 7)

Arguments

x	a stochastic vector or a row-stochastic matrix.
digits	number of digits for rounding.

Details

Rounds and adjusts one entry such that the rounding error is the smallest.

Value

The rounded vector or matrix.

See Also

round

Examples

# regular rounding would not sum up to 1 
x <- c(0.333, 0.334, 0.333)

round_stochastic(x)
round_stochastic(x, digits = 2)
round_stochastic(x, digits = 1)
round_stochastic(x, digits = 0)


# round a stochastic matrix
m <- matrix(runif(15), ncol = 3)
m <- sweep(m, 1, rowSums(m), "/")

m
round_stochastic(m, digits = 2)
round_stochastic(m, digits = 1)
round_stochastic(m, digits = 0)

pomdp documentation built on Sept. 9, 2023, 1:07 a.m.