.. DO NOT EDIT.
.. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY.
.. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE:
.. "info/auto_examples/plot_new_algo.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_algo.py>`
        to download the full example code

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

.. _sphx_glr_info_auto_examples_plot_new_algo.py:


New Algorithm
=====================
A (dummy) example to implement a new PyXAB algorithm. First import all the useful packages

.. GENERATED FROM PYTHON SOURCE LINES 7-14

.. code-block:: default


    from PyXAB.synthetic_obj.Garland import Garland
    from PyXAB.algos.Algo import Algorithm
    from PyXAB.partition.BinaryPartition import BinaryPartition
    import numpy as np









.. GENERATED FROM PYTHON SOURCE LINES 15-17

Now let us suppose that we want to implement a new (dummy) algorithm that always select all nodes on each layer
and then evaluate them for a number of times

.. GENERATED FROM PYTHON SOURCE LINES 17-55

.. code-block:: default


    class Dummy(Algorithm):
        def __init__(self, evaluation=5, rounds=1000, domain=None, partition=None):
            super(Dummy, self).__init__()
            if domain is None:
                raise ValueError("Parameter space is not given.")
            if partition is None:
                raise ValueError("Partition of the parameter space is not given.")
            self.partition = partition(domain=domain)
            self.iteration = 0
            self.evaluation = evaluation
            self.rounds = rounds
            self.partition.deepen()

            self.iterator = 0

        # we need to re-write the pull function
        def pull(self, time):

            depth = self.partition.get_depth()
            index = self.iterator // self.evaluation
            nodes = self.partition.get_node_list()
            point = nodes[depth][index].get_cpoint()
            self.iterator += 1                  # get a point and increase the iterator
            if self.iterator >= self.evaluation * len(nodes[depth]):     # If every point is evaluated, deepen the partition
                self.partition.deepen()

            return point

        # we need to re-write the receive_reward function
        def receive_reward(self, time, reward):
            # No update given the reward for the dummy algorithm
            pass

        def get_last_point(self):

            return self.pull(0)








.. GENERATED FROM PYTHON SOURCE LINES 56-57

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

.. GENERATED FROM PYTHON SOURCE LINES 57-64

.. code-block:: default

    T = 100
    target = Garland()
    domain = [[0, 1]]
    partition = BinaryPartition
    algo = Dummy(rounds=T, domain=domain, partition=partition)        # The new algorithm









.. GENERATED FROM PYTHON SOURCE LINES 65-66

As shown below, the Dummy algorithm now optimizes the objective

.. GENERATED FROM PYTHON SOURCE LINES 66-72

.. 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)
        #print(point)







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

   **Total running time of the script:** (0 minutes 0.004 seconds)


.. _sphx_glr_download_info_auto_examples_plot_new_algo.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_algo.py <plot_new_algo.py>`

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

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


.. only:: html

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

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