Qmkpy

Added
- Add new `json` strategy to save and load problem instances
- Add new (optional) `name` attribute to the `QMKProblem` class
- Add custom `__str__` method to the `QMKProblem` class which prints the name
attribute, if it exists, or a string with the number of items and knapsacks
otherwise.

Updated
- The `io.save_problem_txt` will now first try to use the `problem.name`
atrribute if the `name` keyword is not given.
- Change the (sub)title of the package to "QMKPy: A Python Testbed for the
Quadratic Multiple Knapsack Problem"

Added
- Add example scripts to repository
- Add new utility functions
- `util.get_unassigned_items`
- `util.get_empty_knapsacks`
- `util.get_remaining_capacities`
- Add new check function
- `checks.is_symmetric_profits`
- Update documentation about datasets and add example

Fixed
- Fix docstrings

Added
- Class `QMKProblem` for defining a QMKP
- Functions to save and load QMKProblem instances. The available strategies are
- `numpy` (using numpy's npz format)
- `pickle` (using Python's pickle library)
- `txt` (using a text-based format)
- Module with some checks for feasibility of solutions, if an array is binary
- Some util functions, e.g., to convert the binary assignment matrix to the
chromosome representation
- Solution algorithms for the QMKP
- Constructive procedure
- FCS procedure
- Simple round robin scheme
- Random assignment