sample_mdp.m

来自「approximate reinforcement learning」· M 代码 · 共 20 行

function [xplus, rplus] = sample_mdp(m, x, u)
%  Implements the discrete-time dynamics of the Markov decision process.
%  [XPLUS, RPLUS] = DOUBLEINT_MDP(M, X, U)
%  Parameters:
%   M   - the model specification. Typically contains the fields (all structures)
%           phys - physical parameters
%           disc - discretization configuration
%           goal - goal configuration
%       but the actual structure may depend on the particular MDP.
%   X   - current state, x(k)
%   U 	- command u(k)
%  Returns:
%   XPLUS       - state at next sample, x(k+1)
%   RPLUS       - ensuing reward, r(k+1)

% compute here the next state and reward
xplus = 0 * m.phys.a * x + 0 * u;
rplus = 0;

% END sample_mdp() RETURNING xplus, rplus ===============================================