Home > marl > world > world_dynamics.m

world_dynamics

PURPOSE ^

Template. Implements the discrete-time dynamics of the world

SYNOPSIS ^

function [w, stateviews, actionviews, rewardviews, finished] = world_dynamics(w, actions)

DESCRIPTION ^

Template. Implements the discrete-time dynamics of the world
  [W, STATEVIEWS, ACTIONVIEWS, REWARDVIEWS[, FINISHED]] = WORLD_DYNAMICS(W, ACTIONS)
  Transforms the state of the world based on the agent actions. Also
  computes the instantaneous performance index, the reward function.
  Complete, incomplete or uncertain observations of states, actions, and
  rewards may be implemented via the views mechanism.

  This function is a template and performs no operations.
  A new world type must implement this function under the name
  <world_type>_advance. This functions' input and return arguments must
  conform to the specifications below.
  
  Parameters:
   W           - the world.
   ACTIONS     - the joint action vector of the agents. 
               May contain NaN elements, signifying 'noop' for agents that
               completed their task.

  Returns:
   W           - the world, altered as a result of the agents' actions
   STATEVIEWS  - the views of the agents over the state, nth agent view on
               the nth column.
   ACTIONVIEWS - the views of the agents over the joint action, nth agent
               view on the nth column.
   REWARDVIEWS - the views of the agents over the joint reward, nth agent
               view on the nth column.
   FINISHED    - (only for episodic worlds) a boolean row vector
               containing for each agent a flag that signals if that
               agents' task has finished in this trial (either by success,
               or failure). nth agent flag on the nth column.


  See also world

CROSS-REFERENCE INFORMATION ^

This function calls: This function is called by:
Generated on Wed 04-Aug-2010 16:55:08 by m2html © 2005