Applying self-adaptive evolutionary algorithms to two-dimensional packing problems using a four corners’ heuristic |
| |
Authors: | Kevin J. Binkley Masafumi Hagiwara |
| |
Affiliation: | Department of Information and Computer Science, Hagiwara Research Lab, Keio University, 3-14-1 Hyoshi, Kohoku-ku, Yokohama 223-8522, Japan |
| |
Abstract: | This paper proposes a four corners’ heuristic for application in evolutionary algorithms (EAs) applied to two-dimensional packing problems. The four corners’ (FC) heuristic is specifically designed to increase the search efficiency of EAs. Experiments with the FC heuristic are conducted on 31 problems from the literature both with rotations permitted and without rotations permitted, using two different EA algorithms: a self-adaptive parallel recombinative simulated annealing (PRSA) algorithm, and a self-adaptive genetic algorithm (GA). Results on bin packing problems yield the smallest trim losses we have seen in the published literature; with the FC heuristic, zero trim loss was achieved on problems of up to 97 rectangles. A comparison of the self-adaptive GA to fixed-parameter GAs is presented and the benefits of self-adaption are highlighted. |
| |
Keywords: | Packing Cutting Genetic algorithms Simulated annealing Evolutionary algorithms |
本文献已被 ScienceDirect 等数据库收录! |
|