Speaker
Ms
Junjie Yang
(National University of Defense Technology)
Description
Layout design of spacecraft module belongs to scheme design problem, which has been proved to be NP hard. This problem has not only computing complexity but also engineering complexity, and it is more difficult to tackle the challenge of practical application in engineering. In practical engineering, the dimension of satellite configuration is usually unknown and needs to be optimized (generally minimized) as well, while the dimensions of satellite components are known. Assumed that the material densities of all modules are same, the layout design of satellite module can be simplified to a 2D open dimension rectangle packing problem when the satellite configuration is cuboid. In this problem, given a set of rectangles with known dimensions, the arrangement of these rectangles should be determined without overlapping and inside a predefined enveloping rectangle. This paper proposed a self-boundary fall free genetic algorithm (GA&SBFFA) for the open dimension rectangle packing problem, which is used to optimally arrange the rectangles densely and minimize the area of enveloping rectangle. Meanwhile, the shape of the enveloping rectangle is maintained as square as possible so as to satisfy the static equilibrium requirement in some complex system design, e.g. satellite. The main procedure of SBFFA is as follows. First, the dimension of enveloping rectangle is determined by the packing items, which is different from the methods published previously. Next, the information of feasible space where the next item can be placed is recorded, which includes the widths of the feasible spaces and the coordinates of the feasible points where the bottom left vertex of the next item can be placed. Finally, the next item is placed according to the principle of minimal potential energy. Based on SBFFA, the minimum enveloping rectangle space is calculated with given sequence of packing items. GA is used to solve the packing optimization problem by searching the optimal item sequence. Two experiments are used to testify the proposed method and the efficacy is demonstrated. The computational expenses were reduced to about 30 seconds when there are 50 items, which is much less than the reported methods.
Keywords: Packing problem; Layout optimization; Area minimization.
Layout design of spacecraft module belongs to scheme design problem, which has been proved to be NP hard. This problem has not only computing complexity but also engineering complexity, and it is more difficult to tackle the challenge of practical application in engineering. In practical engineering, the dimension of satellite configuration is usually unknown and needs to be optimized (generally minimized) as well, while the dimensions of satellite components are known. Assumed that the material densities of all modules are same, the layout design of satellite module can be simplified to a 2D open dimension rectangle packing problem when the satellite configuration is cuboid. In this problem, given a set of rectangles with known dimensions, the arrangement of these rectangles should be determined without overlapping and inside a predefined enveloping rectangle. This paper proposed a self-boundary fall free genetic algorithm (GA&SBFFA) for the open dimension rectangle packing problem, which is used to optimally arrange the rectangles densely and minimize the area of enveloping rectangle. Meanwhile, the shape of the enveloping rectangle is maintained as square as possible so as to satisfy the static equilibrium requirement in some complex system design, e.g. satellite. The main procedure of SBFFA is as follows. First, the dimension of enveloping rectangle is determined by the packing items, which is different from the methods published previously. Next, the information of feasible space where the next item can be placed is recorded, which includes the widths of the feasible space and the coordinates of the feasible points where the bottom left vertex of the next item can be placed. Finally, the next item is placed according to the principle of minimal potential energy. Based on SBFFA, the minimum enveloping rectangle space is calculated with given sequence of packing items. GA is used to solve the packing optimization problem by searching the optimal item sequence. Two experiments are used to testify the proposed method and the efficacy is demonstrated. The computational expenses were reduced to about 30 seconds when there are 50 items, which is much less than the reported methods.
Keywords: Packing problem; Layout optimization; Area minimization.
| Applicant type | First author |
|---|
Primary author
Ms
Junjie Yang
(National University of Defense Technology)
Co-authors
Mr
Jie Qi
(National University of Defense Technology)
Mr
Ning Wang
(National University of Defense Technology)
Mr
Xiaoqian Chen
(National University of Defense Technology)
