Retour à la vue des calendrier
Jeudi 7 Avril
Heure: 11:30 - 12:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Quantum Computing for Process Systems Engineering
Description: David Bernal Neira Optimization problems arise in different areas of Process Systems Engineering (PSE), and solving these problems efficiently is essential for addressing important industrial applications.

Quantum computers have the potential to efficiently solve challenging nonlinear and combinatorial problems. However, available quantum computers cannot solve practical problems; they are limited to small sizes and do not handle constraints well. In this talk, we propose hybrid classical-quantum algorithms to solve mixed-integer nonlinear problems (MINLP) and apply decomposition strategies to break down MINLPs into Quadratic Unconstrained Binary Optimization (QUBO) subproblems that can be solved by quantum computers. We will also cover different approaches to solving Quadratic Unconstrained Binary Optimization (QUBO) problems through unconventional computation methods, including but not limited to Quantum algorithms, and discuss how these approaches lead to algorithms able to outperform classical solution approaches
Jeudi 21 Avril
Heure: 10:30 - 11:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Fast algorithms for some parametric optimization problems
Description: Hassan Aissi Parametric optimization is a rich field with applications ranging from sensitivity analysis, Lagrangian relaxation, multiobjective optimization, and minimum-ratio optimization. We consider in this talk some parametric problems related to the minimum cut, in which we are given a graph G=(V,E) with edge costs that are affine functions of a parameter ???d. We develop strongly polynomial algorithms for these problems that are faster than known techniques.
Jeudi 2 Juin
Heure: 11:45 - 12:30
Lieu: Visio - https://bbb.lipn.univ-paris13.fr/b/wol-ma9-vjn -code 514019
Résumé: On two two-level problems for operational warehouse planning in person-to-parts order picking systems
Description: Stefan Irnich We present a new modeling and solution approach for two-level problems in warehousing where one level concerns picking operations in a manual picker-to-parts warehouse. In particular, we consider the single picker routing problem with scattered storage (SPRP-SS) and the joint order batching and picker routing problem (JOBPRP). The SPRP-SS assumes that an article is, in general, stored at more than one pick position. The task is then the simultaneous selection of pick positions for requested articles and the determination of a minimum-length picker tour collecting the articles. In the JOBPRP, a set of orders is given, each with one or several order lines requesting a number of articles. The problem is here to group the given orders into capacity-feasible batches so that the total length of the picker tours collecting the respective articles is minimized. It is a classical result of Ratliff and Rosenthal that, for given pick positions, an optimal picker tour is a shortest path in the state space of a dynamic program with a linear number of states and transitions. We extend the state space of Ratliff and Rosenthal so that every feasible picker tour is still a path. Furthermore, the additional requirement to make consistent selections and grouping decisions can be modeled as additional constraints in shortest-path problems. We propose to solve these problems with a MIP solver. We will explain why this approach is not only convenient and elegant but also generic: it covers optimal solutions to integrated problems that use heuristic routing policies for the picker tours, consider different warehouse layouts, and incorporate further extensions. Computational experiments with a direct MIP solver-based approach for the SPRP-SS and a branch-price-and-cut algorithm for the JOBPRP show that the new modeling and solution approach outperforms the available exact algorithms. The latter computes hundreds of new best and provably optimal solutions to open instances of three JOBPRP benchmark sets. (joint work with Katrin Heßler)