pyepo.model.grb.tsp
===================

.. py:module:: pyepo.model.grb.tsp

.. autoapi-nested-parse::

   Traveling salesman problem



Classes
-------

.. autoapisummary::

   pyepo.model.grb.tsp.tspABModel
   pyepo.model.grb.tsp.tspGGModel
   pyepo.model.grb.tsp.tspGGModelRel
   pyepo.model.grb.tsp.tspDFJModel
   pyepo.model.grb.tsp.tspMTZModel
   pyepo.model.grb.tsp.tspMTZModelRel


Module Contents
---------------

.. py:class:: tspABModel(num_nodes: int, *args, **kwargs)

   Bases: :py:obj:`pyepo.model.bases.tspABBase`, :py:obj:`pyepo.model.grb.grbmodel.optGrbModel`


   Gurobi-backed TSP abstract base. Provides paired-variable ``setObj`` /
   ``solve`` / ``_addExtraConstr`` shared by GG and MTZ. DFJ overrides those.


   .. py:method:: setObj(c: numpy.ndarray | torch.Tensor | list) -> None

      A method to set the objective function

      :param c: cost vector



   .. py:method:: solve() -> tuple[numpy.ndarray, float]

      A method to solve model



.. py:class:: tspGGModel(num_nodes: int, *args, **kwargs)

   Bases: :py:obj:`tspABModel`


   Gavish-Graves (GG) formulation.


   .. py:method:: relax() -> tspGGModelRel

      A method to get linear relaxation model



.. py:class:: tspGGModelRel(num_nodes: int, *args, **kwargs)

   Bases: :py:obj:`tspGGModel`


   LP relaxation of the GG formulation.


   .. py:method:: solve() -> tuple[numpy.ndarray, float]

      A method to solve model — returns fractional solution.



   .. py:method:: relax() -> NoReturn

      A forbidden method to relax MIP model



   .. py:method:: getTour(sol: numpy.ndarray | torch.Tensor | list) -> list[int]

      A forbidden method to get a tour from solution



.. py:class:: tspDFJModel(num_nodes: int, recycle_cuts: bool = False, *args, **kwargs)

   Bases: :py:obj:`tspABModel`


   Danzig-Fulkerson-Johnson (DFJ) formulation with lazy subtour elimination.

   Uses one undirected Var per edge (``x[j,i]`` aliases ``x[i,j]``), so the
   paired-vars ``setObj`` / ``solve`` / ``_addExtraConstr`` inherited from
   ``tspABModel`` are overridden here. With ``recycle_cuts`` the subtour cuts
   found in one solve join the model's lazy pool, so later solves on new cost
   vectors skip rediscovering them.


   .. py:attribute:: recycle_cuts
      :value: False



   .. py:method:: setObj(c: numpy.ndarray | torch.Tensor | list) -> None

      A method to set the objective function

      :param c: cost vector



   .. py:method:: solve() -> tuple[numpy.ndarray, float]

      A method to solve model



.. py:class:: tspMTZModel(num_nodes: int, *args, **kwargs)

   Bases: :py:obj:`tspABModel`


   Miller-Tucker-Zemlin (MTZ) formulation.


   .. py:method:: relax() -> tspMTZModelRel

      A method to get linear relaxation model



.. py:class:: tspMTZModelRel(num_nodes: int, *args, **kwargs)

   Bases: :py:obj:`tspMTZModel`


   LP relaxation of the MTZ formulation.


   .. py:method:: solve() -> tuple[numpy.ndarray, float]

      A method to solve model — returns fractional solution.



   .. py:method:: relax() -> NoReturn

      A forbidden method to relax MIP model



   .. py:method:: getTour(sol: numpy.ndarray | torch.Tensor | list) -> list[int]

      A forbidden method to get a tour from solution