Triangulation¶

class safe_learning.Triangulation(discretization, vertex_values, project=False, name='triangulation')¶

Efficient Delaunay triangulation on regular grid.

This is a tensorflow wrapper around a numpy implementation.

This class is a wrapper around scipy.spatial.Delaunay for regular grids. It splits the space into regular hyperrectangles and then computes a Delaunay triangulation for only one of them. This single triangulation is then generalized to other hyperrectangles, without ever maintaining the full triangulation for all individual hyperrectangles.

Parameters:	discretization : instance of discretization For example, an instance of GridWorld. vertex_values : arraylike, optional A 2D array with the values at the vertices of the grid on each row. Is converted into a tensorflow variable. project : bool, optional Whether to project points onto the limits. name : string The tensorflow scope for all methods.
Attributes:	`discretization` Getter for the discretization. `nindex` Return the number of parameters. `parameters` Return the variables within the current scope. `project` Getter for the project parameter. scope_name

Methods

`__call__`(self, \args, \\*kwargs)	Evaluate the function using the template to ensure variable sharing.
`build_evaluation`(self, points)	Evaluate using tensorflow.
`copy_parameters`(self, other_instance)	Copy over the parameters of another instance.
`gradient`(self, points)	Compute derivatives using tensorflow.

build_evaluation(self, points)¶: Evaluate using tensorflow.

copy_parameters(self, other_instance)¶: Copy over the parameters of another instance.

discretization¶: Getter for the discretization.

gradient(self, points)¶: Compute derivatives using tensorflow.

nindex¶: Return the number of parameters.

parameters¶: Return the variables within the current scope.

project¶: Getter for the project parameter.