DagSemProc.04461.11.pdf
- Filesize: 1.05 MB
- 15 pages
Document
Hybrid Representations for Composition Optimization and Parallelizing MOEAs
Authors
-
Part of:
Volume:
Dagstuhl Seminar Proceedings, Volume 4461
Series: Dagstuhl Seminar Proceedings (DagSemProc) - License:
Creative Commons Attribution 4.0 International license
- Publication Date: 2005-08-10
Cite AsGet BibTex
Felix Streichert, Holger Ulmer, and Andreas Zell. Hybrid Representations for Composition Optimization and Parallelizing MOEAs. In Practical Approaches to Multi-Objective Optimization. Dagstuhl Seminar Proceedings, Volume 4461, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2005)
https://doi.org/10.4230/DagSemProc.04461.11
https://doi.org/10.4230/DagSemProc.04461.11
Abstract
We present a hybrid EA representation suitable to optimize composition optimization problems ranging from optimizing recipes for catalytic materials to cardinality constrained portfolio selection. On several problem instances we can show that this new representation performs better than standard repair mechanisms with Lamarckism.
Additionally, we investigate the a clustering based parallelization scheme for MOEAs. We prove that typical "divide and conquer'' approaches are not suitable for the standard test functions like ZDT 1-6. Therefore, we suggest a new test function based on the portfolio selection problem and prove the feasibility of "divide and conquer'' approaches on this test function.
Keywords
- Multi-objective Evolutionary Algorithms (MOEAs)
- Solution Representation
- Constrained Portfolio Selection Problem
- Parallelizing MOEAs
Metrics
- Access Statistics
-
Total Accesses (updated on a weekly basis)
0PDF Downloads