Communication Dans Un Congrès Année : 2024

Bi-objective optimization for reconfigurable assembly line design under uncertainty

Résumé

Reconfigurable manufacturing systems can adjust their structure to respond to the functionalities and capabilities required for a product family. By integrating the concept of reconfigurability into traditional assembly lines, it becomes possible to add extra manufacturing resources to each station. The assembly line’s performance mainly depends on task assignment. However, continuously reassigning tasks for production adjustments is costly and not appreciated by manufacturers. The alternative, more convenient approach is to maintain line balancing while modifying resource allocation in order to generate different configurations. Adding one resource to a station could either improve the productivity of the assembly line, or enhance its robustness when facing uncertain process times. Hence, we introduce a sequential process that starts with defined line balancing and generates configurations by gradually adding resources until reaching the maximum number of available resources. The only interesting configurations are those that provide a suitable trade-off between productivity and robustness for the same number of resources, following the principle of Pareto optimality. To evaluate the quality of these trade-offs, the pervolume metric, is employed. A simulated annealing algorithm is developed to determine the balancing with the highest hypervolume. Numerical experiments are conducted to test the algorithm. Computational results will be presented at the conference.

Fichier non déposé

Dates et versions

hal-04925787 , version 1 (02-02-2025)

Identifiants

  • HAL Id : hal-04925787 , version 1

Citer

Siwar Arbi, Audrey Cerqueus, Alexandre Dolgui, Evgeny Gurevsky, Ali Siadat. Bi-objective optimization for reconfigurable assembly line design under uncertainty. Conference on Optimization, Analytics, and Decisions in the Big Data Era, in honor of the 70th birthday of Professor Panos M. Pardalos, Jun 2024, Halkidiki, Greece. ⟨hal-04925787⟩
136 Consultations
0 Téléchargements

Partager

  • More