|
Authors: | Nando Laumanns, Marco Laumanns, Dirk Neunzig |
Group: | Computer Engineering |
Type: | Inproceedings |
Title: | Multi-Objective Design Space Exploration of Road Trains with Evolutionary Algorithms |
Year: | 2001 |
Month: | March |
Pub-Key: | LLN2001a |
Book Titel: | Lecture Notes on Computer Science. Evolutionary Multi-criterion Optimization (EMO 2001) |
Volume: | 1993 |
Pages: | 612-623 |
Keywords: | EMO DSE |
Publisher: | Springer |
Abstract: | This paper examines the road train concept as a new alternative in long-distance freight traffic. The design of such a system is a difficult task since many different and conflicting criteria arise depending on the application spec-trum, the legal conditions and the preferences of the carrier. Furthermore the evaluation of each decision alternative relies on a time consuming and sophisti-cated simulation. Evolutionary algorithms (EAs) have shown to be a useful tool for multi-objective optimization in engineering design. Based on a unified model, we develop a problem-specific evolutionary algorithm which features strong elitism, an unlimited archive of non-dominated solutions and density de-pendent selection. This EA is able to create alternatives which dominate previ-ous manually engineered solutions as well as those derived from exhaustive search. |
Location: | Berlin |
Resources: | [BibTeX] [Paper as PDF] |