rcss (version 1.1)

StochasticGrid: Stochastic grid

Description

Generate a grid using k-means clsutering.

Usage

StochasticGrid(start, disturb, n_grid, max_iter, warning)

Arguments

start

Array representing the start. The first entry must be 1 and array [-1] represents the starting state.

disturb

4 dimensional array containing the path disturbances. Matrix [,,i,j] represents the disturbance at time i for sample path j.

n_grid

Number of grid points in the stochastic grid.

max_iter

Maximum iterations in the k-means clustering algorithm.

warning

Boolean indicating whether messages from the k-means clustering algorithm are to be displayed

Value

Matrix representing the stochastic matrix. Each row represents a particular grid point. The first column contains only 1s.

Examples

## Generate a stochastic matrix using an AR(2) process
start <- c(1, 0, 0)
n_dim <- length(start)
n_path <- 10
psi1 <- 0.3
psi2 <- 0.65
n_dec <- 21
path_disturb <- array(data = matrix(c(1, 0,    0,
                                      0, 0,    1,
                                      0, psi2, psi1), ncol = 3, byrow = TRUE),
                      dim = c(n_dim, n_dim, (n_dec - 1), n_path))
path_disturb[3,1,,] <- runif((n_dec - 1) * n_path, -1, 1)
n_grid <- 10
grid <- StochasticGrid(start, path_disturb, n_grid, 10, TRUE)