fontanf / teaching

I develop and maintain:

• A solver dedicated to geometrical packing problems fontanf/packingsolver

 

 

• A solver for some timetabling problems fontanf/timetablingsolver

   

• Packages that simplify the implementation of combinatorial optimization algorithms based on the following methods:
  • Local search fontanf/localsearchsolver (C++) fontanf/localsearchsolverpy (Python3)
  • Heuristic tree search fontanf/treesearchsolver (C++) fontanf/treesearchsolverpy (Python3)
  • Column generation heuristics fontanf/columngenerationsolver (C++) fontanf/columngenerationsolverpy (Python3)

   

• Solvers for various fundamental operations research problems:
  • 0-1 knapsack problem, subset-sum problem, multiple-choice subset-sum problem fontanf/knapsacksolver
  • Generalized assignment problem fontanf/generalizedassignmentsolver
  • (Unicost) set covering problem fontanf/setcoveringsolver
  • Maximum(-weight) independent set and maximum-(weight) clique problems fontanf/stablesolver
  • Graph coloring problem fontanf/coloringsolver
  • Traveling salesman problem fontanf/travelingsalesmansolver

• Solvers for some other operations research problems:

  • Star observation scheduling problems fontanf/starobservationschedulingsolver
  • Travelling thief problems fontanf/travellingthiefsolver

• A repository containing data, instance parsers and solution checkers for various operations research problems fontanf/orproblems. These problems are used in the examples of the packages mentioned above.

• A repository containing various tools used by my other repositories fontanf/optimizationtools

Finally, the repository fontanf/teaching contains the material I use to teach.

Do not hesitate to let me know if you use some of my codes. I'm happy to know that they are useful. And I'll be more careful not to break everything in the next update. And do not hesitate to give me some feedback as well.

teaching's Issues

Problem when instance solution is the empty subset.

In the batch scheduling project, when we try to check a certificate with an empty set of items we get an error because we are computing the max over an empty set. This happens for example in the first instance where we have no items (and solution with no items).
It's not clear from the subject what the maximum width means when no items are selected. Is it zero?
In this case adding the argument default=0 to the the call of the max function should fix this:

width = max((self.items[item_id].width
             for item_id in data["items"]), default=0)