value_iteration: Identify the dynamic optimum using backward iteration...
In cboettig/pdg_control: Pretty Darn Good Control

Identify the dynamic optimum using backward iteration (dynamic programming)

1 2	value_iteration(SDP_Mat, x_grid, h_grid, OptTime = 100, xT, profit, delta, epsilon = 1e-04)

`SDP_Mat`	the stochastic transition matrix at each h value
`x_grid`	the discrete values allowed for the population size, x
`h_grid`	the discrete values of harvest levels to optimize over
`OptTime`	the stopping time
`xT`	the boundary condition population size at OptTime
`profit`	the profit function (i.e. enforces the boundary condition)
`delta`	the discounting rate (1-delta)
`epsilon`	value iteration tolerance

list containing the matrices D and V. D is an x_grid by OptTime matrix with the indices of h_grid giving the optimal h at each value x as the columns, with a column for each time. V is a matrix of x_grid by x_grid, which is used to store the value function at each point along the grid at each point in time. The returned V gives the value matrix at the first (last) time.

cboettig/pdg_control documentation built on May 13, 2019, 2:10 p.m.