1: \begin{abstract}

2:   Propagation of linear constraints has become a crucial sub-routine in

3: modern Mixed-Integer Programming (MIP) solvers.

4: %

5: In practice, iterative algorithms with tolerance-based stopping criteria are

6: used to avoid problems with slow or infinite convergence.

7: %

8: However, these heuristic stopping criteria can pose difficulties for fairly

9: comparing the efficiency of different implementations of iterative

10: propagation algorithms in a real-world setting.

11: %

12: Most significantly, the presence of unbounded variable domains in the problem 

13: formulation makes it difficult to quantify the relative size of reductions performed on them.

14: %

15: In this work, we develop a method to measure---independently of the algorithmic

16: design---the progress that a given iterative propagation procedure has

17: made at a given point in time during its execution.

18: %

19: Our measure makes it possible to study and better compare the behavior of bounds

20: propagation algorithms for linear constraints.

21: %

22: We apply the new measure to answer two questions of practical relevance:

23: %

24: (i) We investigate to what extent heuristic stopping criteria can lead to

25: premature termination on real-world MIP instances.

26: %

27: (ii) We compare a GPU-parallel propagation algorithm against a sequential

28: state-of-the-art implementation and show that the parallel version is even more

29: competitive in a real-world setting than originally reported.

30: \end{abstract}

31: