The behavior of LP models is a topic in which I have had some interest in recent years. I speak of behavior of the model itself, as distinct from the behavior of the algorithm, the solution path or the accuracy of the model representation. I have submitted several papers on this topic before to the SIGMAP bulletin (1, 2, 3). In this paper I consider model behavior as reflected in alternate optimal solutions.
/lp/association-for-computing-machinery/alternate-optimal-solutions-CSfVRz0lrV