- when doing a limited-depth backtracking, but no suggestion is found
- only do a random assignment with >1 conflicts when Suggestion.AllowUnassignments is true (defaults to false)
- while this somewhat limits the Suggestion's ability to converge the search towards a complete solution, allowing unassignments (>1 conflicts) causes the Hill Climber to diverge from a local optima
- this is because the number of assigned variables is not included in the objective function that is being minimized
- other search algorithms (such as Great Deluge or Simulated Annealing) can also suffer from accepting a move that decreases the number of assigned variables, especially when it is not possible to find a complete solution eventually (i.e., no complete solution can be found)
