MP or not MP: that is the question

It is well known that in the twentieth century, mathematical programming (MP) modeling and particularly linear programming (LP) modeling, even though strongly applied to combinatorial optimization, were not too successful when directed to scheduling problems. The purpose of this paper is to show that the field of successful applications of LP/MP modeling is still growing and includes also scheduling topics. We first focus on single machine scheduling. We consider a single machine scheduling model where a quadratic programming (QP) formulation handled by means of a QP solver is shown to be competitive with the state of the art approaches. Also, we discuss a single machine bicriterion scheduling problem and show that a standard LP formulation based on positional completion times performs reasonably well when handled by means of a LP solver. Then, we show how LP can be used to tighten bounds for approximation results in sequencing problems. Finally, we show how to enhance the complexity bounds of branch-and-reduce exact exponential algorithms by means of the so-called measure-and-conquer paradigm requiring always the solution of a specific MP model.