The management of transportation and production systems often requires solving a sequence of optimization problems, each problem optimizing the utilization of some resources: equipment, personnel, etc.
Dr. François Soumis, (École Polytechnique de Montréal)
The management of transportation and production systems often requires solving a sequence of optimization problems, each problem optimizing the utilization of some resources: equipment, personnel, etc. For instance, transit authorities perform bus scheduling followed by daily and monthly driver scheduling; airlines perform aircraft scheduling followed by crew pairing and monthly crew scheduling; and manufacturing companies address manpower scheduling before production scheduling. Such a sequential approach for management was introduced a long time ago when solutions were computed manually. A large improvement in productivity was obtained by replacing manual decision processes by optimization models and methods. However, the global decision process is not yet optimal since optimal decisions for the first problems of a sequence of problems can reduce the feasible domain of the subsequent problems and increase drastically the cost of their solutions. The objective of this project is to develop the know-how for optimizing simultaneously several problems in such sequences. It is a challenge involving a reengineering of the decision processes and requesting the solution of very large scale combined optimization problems.