Cluster Events
NEXT SEMINAR 2024/2025
In this section, you will find the information (Speaker/Guest, abstract, date & time slot , location, online registration link..) about the upcoming seminar during the year. Everyone is welcome, from within ESSEC as well as from outside.
Axel PARMENTIER - ENPC Ponts et Chaussées
Tuesday, March 18th 2025 from 12:15 pm to 1pm
Room : N230
Zoom link : https://essec.zoom.us/j/99754570811
Title: Continental Scale Inventory Routing at Renault.
Abstract : Renault transports car components from suppliers to plants using specific packaging to prevent damage. To reduce return logistics costs, Renault shares a pool of standardized packaging among all suppliers. However, pooling introduces complexity to return logistics, creating an inventory routing problem on a much larger scale than typically studied. This talk addresses two aspects of the problem: the operational setting, which involves daily decisions on packaging distribution and route planning, and the tactical problem of negotiating three-month contracts with carriers. The operational setting is tackled using Bayesian machine learning and a large neighborhood search. The algorithm is now in production, leading to significant cost and CO2 emissions reductions. For the tactical problem, combinatorial optimization augmented with machine learning is used to propose relevant contracts, relying on recent advances in empirical risk minimization in structured learning.
****
Marianna DE SANTIS - Università degli studi di Firenze
Tuesday, March 11st 2025 from 12pm to 1pm
Room : N406
Zoom link : https://essec.zoom.us/j/92588571771
Title: Challenges in solving multiobjective mixed integer nonlinear programming: computation of efficient integer assignments and enclosures
Abstract: Multiobjective mixed integer nonlinear programming (MOMINLP) represents a powerful tool to model real-world decision problems, as applied problems often reveal multiple, conflicting goals to be optimized and 0-1 or integer variables are needed to model logical relationships or discrete quantities. Solving MOMINLPs means to detect the set of efficient solutions of the problem, i.e. those points such that none of the objective functions can be improved in value without degrading some of the other objective values.
An efficient integer assignment for MOMINLPs is a fixing of the integer variables so that there exists an efficient point of the problem with exactly that fixing. Depending on the instance, solution algorithms for MOMINLPs may need to detect a large number of efficient integer assignments, possibly all integer feasible assignments, so that a full enumeration cannot be avoided. This shows a big difference with respect to the single objective case and a big challenge in terms of algorithms development.
In this talk, we present a branch-and-bound method for multiobjective mixed-integer convex quadratic programs that computes a superset of efficient integer assignments using outer approximations of the upper image set of continuous relaxations. These outer approximations are obtained addressing the dual formulations of specific subproblems where the values of certain integer variables are fixed. The devised pruning conditions and a tailored preprocessing phase allow a fast enumeration of the nodes. Despite the fact that we do not require any boundedness of the feasible set, we are able to prove that the method stops after having explored a finite number of nodes. Numerical experiments on instances with two, three, and four objectives are presented.
If you need more information, please contact matta@essec.edu
Best regards,