Bin-packing problems consist in determining how many objects of predetermined forms can be placed in a certain volume or a certain surface as well as the optimal configuration of those.
To solve this problem, one usually transpose it to an equivalent optimization problem. This generally admits many local minima.
This report aims solving this problem by using stochastic-based optimization methods, more particularly certain evolutionary algorithms used in the context of global non-linear optimization, introduced by Catherine Roegiers into her report.
Once this kind of algorithms is introduced, the successive stages required to solve the simple bin-packing problem (determining the maximum number of discs that can be placed in a rectangular container) will be presented.
Utilisation d'algorithmes évolutionnaires dans le cadre des problèmes d'empaquetage d'objets
Mahieu, O. (Author). 2005
Student thesis: Master types › Master in Mathematics