5. Conclusion

With the help of examples, we have tried to present in this dossier the best-known and most widely used methods and techniques for solving discrete optimization problems. We have made the strong assumption that a single objective function can be used to measure the value of a solution. This assumption is not always acceptable, however, as there may be several conflicting criteria, such as quality, price and lead time. In this case, we can no longer optimize a single objective function, but seek a compromise, known as Pareto optima, between several criteria. The discipline corresponding to this search is called multi-criteria analysis, and has given rise to other, even more complex methods of resolution, usually involving the interactive participation of decision-makers. It should be noted, however, that to find Pareto optima, it is necessary to solve continuous or discrete single-criteria...