.. DO NOT EDIT.
.. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY.
.. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE:
.. "info/auto_examples/plot_new_partition.py"
.. LINE NUMBERS ARE GIVEN BELOW.

.. only:: html

    .. note::
        :class: sphx-glr-download-link-note

        :ref:`Go to the end <sphx_glr_download_info_auto_examples_plot_new_partition.py>`
        to download the full example code

.. rst-class:: sphx-glr-example-title

.. _sphx_glr_info_auto_examples_plot_new_partition.py:


New Partition
=====================
An example to implement a new partition of the space for any PyXAB algorithm. First import all the useful packages

.. GENERATED FROM PYTHON SOURCE LINES 7-17

.. code-block:: default


    from PyXAB.synthetic_obj.Garland import Garland
    from PyXAB.algos.HOO import T_HOO
    from PyXAB.partition.Partition import Partition
    from PyXAB.partition.Node import P_node
    import numpy as np











.. GENERATED FROM PYTHON SOURCE LINES 18-21

Now let us suppose that we want to implement a new binary partition that always split the domain into
two nodes that are 1/3 and 2/3 of its original size, i.e., if the original projection on the chosen dimension is
``[a, b]``, we split the domain into ``[a, 0.67a + 0.33b]`` and ``[0.67a + 0.33b, b]``.

.. GENERATED FROM PYTHON SOURCE LINES 21-75

.. code-block:: default


    class NewBinaryPartition(Partition):

        def __init__(self, domain, node=P_node):

            super(NewBinaryPartition, self).__init__(domain=domain, node=node)

        # We rewrite the make_chilren function for the new partition
        def make_children(self, parent, newlayer=False):

            parent_domain = parent.get_domain()
            dim = np.random.randint(0, len(parent_domain))
            selected_dim = parent_domain[dim]

            domain1 = parent_domain.copy()
            domain2 = parent_domain.copy()

            # New choice of the split point
            split_point = 2/3 * selected_dim[0] + 1/3 * selected_dim[1]             # split point
            domain1[dim] = [selected_dim[0], split_point]
            domain2[dim] = [split_point, selected_dim[1]]

            # Initialization of the two new nodes
            node1 = self.node(
                depth=parent.get_depth() + 1,
                index=2 * parent.get_index() - 1,
                parent=parent,
                domain=domain1,
            )
            node2 = self.node(
                depth=parent.get_depth() + 1,
                index=2 * parent.get_index(),
                parent=parent,
                domain=domain2,
            )

            # Update the children of the parent
            parent.update_children([node1, node2])

            new_deepest = []
            new_deepest.append(node1)
            new_deepest.append(node2)

            # If creating a new layer, use the new nodes as the first nodes in the new layer
            if newlayer:
                self.node_list.append(new_deepest)
                self.depth += 1
            # Else, just append the new nodes to the existing layer
            else:
                self.node_list[parent.get_depth() + 1] += new_deepest











.. GENERATED FROM PYTHON SOURCE LINES 76-77

Define the number of rounds, the target, the domain, the partition, and the algorithm for the learning process

.. GENERATED FROM PYTHON SOURCE LINES 77-84

.. code-block:: default

    T = 100
    target = Garland()
    domain = [[0, 1]]
    partition = NewBinaryPartition                      # the new partition chosen is NewBinaryPartition
    algo = T_HOO(domain=domain, partition=partition)









.. GENERATED FROM PYTHON SOURCE LINES 85-86

As shown below, the partition should be working

.. GENERATED FROM PYTHON SOURCE LINES 86-92

.. code-block:: default


    for t in range(1, T+1):

        point = algo.pull(t)
        reward = target.f(point) + np.random.uniform(-0.1, 0.1)     # uniform noise
        algo.receive_reward(t, reward)








.. rst-class:: sphx-glr-timing

   **Total running time of the script:** (0 minutes 0.093 seconds)


.. _sphx_glr_download_info_auto_examples_plot_new_partition.py:

.. only:: html

  .. container:: sphx-glr-footer sphx-glr-footer-example




    .. container:: sphx-glr-download sphx-glr-download-python

      :download:`Download Python source code: plot_new_partition.py <plot_new_partition.py>`

    .. container:: sphx-glr-download sphx-glr-download-jupyter

      :download:`Download Jupyter notebook: plot_new_partition.ipynb <plot_new_partition.ipynb>`


.. only:: html

 .. rst-class:: sphx-glr-signature

    `Gallery generated by Sphinx-Gallery <https://sphinx-gallery.github.io>`_