Plénières de COSI - 2017
- Modeling and Solving a class of Real-life Bin Packing problems in Operational Supply Chain
- Validation des transformations de modèles: Concepts et Défis
- Graph packing problems: theory and applications
- Fuzzy Dual Numbers and Optimization under Uncertainty
Abderahmane Aggoun .
Allaoua Chaoui.
Hamamache Khedouci.
Felix Mora-Camino.
Modeling and Solving a class of Real-life Bin Packing problems in Operational Supply Chain
KLS OPTIM is an SME specialist in Warehouse Management Systems and Optimization in Logistics, and offers a complete range of proven packages, solutions and services in a niche of sectors in Supply Chain Management. The main problems addressed are packing, design of optimal plan of packing products in cartons and cartons in pallets, optimization in distribution by minimizing the number of pallets, optimization of vehicle/ container loading plans, optimization of assignment of containers in wagons, integration of packing and vehicle routing. Most of the problems are known in the literature as bin packing problems. In this presentation, we are interested in one hand to review a set of problems encountered in logistics, and in a second hand to highlight the contribution of Constraint Programming, Linear Programming, Metaheuristics and numerical solvers using Non-Linear Inequalities in solving them.Abder Aggoun.
KLS OPTIM, France.
Validation des transformations de modèles: Concepts et Défis
L'objectif de cette conférence invitée est de donner les concepts fondamentaux de la validation des transformations, un survey des différents travaux dans ce domaine ainsi que quelques problèmes ouverts dans ce domaine de recherche Allaoua Chaoui.
Université Constantine 2 - Abdelhamid Mehri.
Graph packing problems: theory and applications
In this talk, I review, classify and discuss several known
conjectures, recent advances and results obtained on graph packing
problems. In particular, I give bounds and algorithms for these problems
and some of their applications in computer science. Hamamache Khedouci.
Université Lyon 1, France..
Fuzzy Dual Numbers and Optimization under Uncertainty
In general, optimization problems assume implicitly that their parameters (cost coefficients, limit values) are perfectly known while very often for real problems this is not the case. A first approach is to perform around the nominal optimal solution a post optimization sensibility analysis. When some probabilistic information about the values of the uncertain parameters is available, stochastic optimization techniques may provide the most expected optimal solution. When these parameters are only known to remain within some intervals, robust optimization techniques will provide a robust solution. The fuzzy formalism has been also considered in this case as an intermediate approach to represent the parameter uncertainties and provide fuzzy solutions. These different approaches result in general into a very large amount of computation.
In this talk, a new formalism based on fuzzy dual numbers is proposed to diminish the computational burden when dealing with uncertainty in mathematical programming problems. In this framework, a solution approach is proposed for mathematical programming problems presenting parameter uncertainty and solution diversion.
The adopted formalism considers fuzzy dual numbers which are a simplified version of fuzzy numbers which adopts some elements of dual numbers calculus. The proposed special class of numbers, dual fuzzy numbers integrates the nilpotent operator Epsilon of dual numbers and considers symmetrical fuzzy numbers. Here after introducing the elements of fuzzy dual calculus useful in this study, two classes of fuzzy dual mathematical programming problems are considered: those where uncertainty relays only in the parameters of the problem and those where also the implementation of the solution is subject to uncertainty. Finally fuzzy dual dynamic programming is developed to solve sequential optimization problems under interval uncertainty.
Felix Mora-Camino.
ENAC, Toulouse, France..