Efficient multi-objective simulated annealing algorithm for interactive layout problems - Archive ouverte HAL Access content directly
Journal Articles International Journal on Interactive Design and Manufacturing Year : 2021

Efficient multi-objective simulated annealing algorithm for interactive layout problems

Abstract

This paper presents an efficient simulated annealing algorithm for solving multi-objective layout problems where several rectangular components are placed, respecting non-overlap and non-protrusion constraints in the given space. Resolving layout problems can be very hard in some industrial cases because problems are over-constrained and computing feasible optimal layout designs are time consuming. In most practical problems, both real and virtual components exist. The virtual components represent the required accessible space allowing the user to access to the component. The virtual components can overlap with each other, while the overlap is not allowed for the real components. Considering the limited layout design space, the capacity of the layout problem is analyzed using constructive placing techniques. To explore the feasible layout space, a hybrid simulated annealing is proposed to determine the order of placement; then, a constructive placing strategy based on empty maximal space is developed. What’s more, an interactive visualization environment is introduced between the optimization and expert. The proposed algorithm is the first attempt to search feasible space of multi-objective layout design by constructive placing method.
Fichier principal
Vignette du fichier
hal-03553932.pdf (977.33 Ko) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-03553932 , version 1 (01-03-2022)

Identifiers

Cite

Xiaoxiao Song, Emilie Poirson, Yannick Ravaut, Fouad Bennis. Efficient multi-objective simulated annealing algorithm for interactive layout problems. International Journal on Interactive Design and Manufacturing, 2021, 15 (4), pp.441-451. ⟨10.1007/s12008-021-00773-1⟩. ⟨hal-03553932⟩
58 View
193 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More