Policy

   Policy constructor.

   Defines an empty Policy object.
   A Policy is a set of elements  that can be vectors or Gaussian mixtures 
   depending on the nature of the state space (discrete vs. continuous).
   Each element in a policy has an associated action.
   The element (and thus, the action) to be assigned to each belief is the
   one that maximizes the value for that belief. Recall that the value is
   computed using the expectation operator between the element and the
   belief. See the different belief implementations for details on this
   operator.

0001 function P=Policy(varargin)
0002 %   Policy constructor.
0003 %
0004 %   Defines an empty Policy object.
0005 %   A Policy is a set of elements  that can be vectors or Gaussian mixtures
0006 %   depending on the nature of the state space (discrete vs. continuous).
0007 %   Each element in a policy has an associated action.
0008 %   The element (and thus, the action) to be assigned to each belief is the
0009 %   one that maximizes the value for that belief. Recall that the value is
0010 %   computed using the expectation operator between the element and the
0011 %   belief. See the different belief implementations for details on this
0012 %   operator.
0013 
0014   switch nargin
0015     case 0
0016       P.elements={};
0017       P.actions={};
0018       P.n=0;
0019       P=class(P,'Policy');
0020     case 1
0021       if isa(varargin{1},'Policy')
0022         P=varargin{1};
0023       else
0024         error('Wrong parameter type in Policy copy constructor.');
0025       end
0026     otherwise
0027       error('Wrong number of parameters in Policy creation.');
0028   end