COMPARING MINIMUM NEIGHBORHOOD EVALUATION SCHEMES FOR FINDING SPATIALLY ROBUST SOLUTIONS
Abstract
The common definition for robust solutions considers a solution robust if it remains optimal when the parameters defining the fitness function are perturbed. A second definition that can be found in the literature: robustness occurs when a solution can be varied spatially without a significant drop in fitness. We propose an alternative operational definition for spatial robustness: both the solution and the neighbourhood around the solution has fitness above a given threshold. With this new definition, we created a set of functions with useful properties to allow for the testing of solution robustness. The performance of a Genetic Algorithm (GA) is then evaluated based on its ability to identify multiple robust solutions based on the above robustness definition. Different neighbourhood evaluation schemes are identified from the literature and compared, with the minimum neighbour technique proving to be the most effective.
References
-
K. Hacker and K. Lewis , Robust Design Through the Use of a Hybrid Genetic Algorithm , ASME 2002 Design Engineering Technical Conferences and Computers and Information in Engineering Conference ( 2002 ) . Google Scholar - Ray, T., Constrained Robust Optimal Design using a Multiobjective Evolutionary Algorithm. IEEE, 2002 . Google Scholar
- Sevaux, M. and K. Sörensen, A genetic algorithm for robust schedules. 2002, University of Valenciennes: Valenciennes . Google Scholar
- Jensen, M.T. and T.K. Hansen, Robust solutions to Job Shop problems. 1999, Aarhus University . Google Scholar
-
K. Sörensen , Tabu Searching for Robust Solutions , 4th Metaheuristics International Conference ( 2001 ) . Google Scholar - IEEE Transaction on Evolutionary Computation 9(3), 303 (2005), DOI: 10.1109/TEVC.2005.846356. Crossref, ISI, Google Scholar
- IEEE Transaction on Evolutionary Computation 1(3), 201 (1997), DOI: 10.1109/4235.661550. Crossref, Google Scholar
-
Y. Jin and B. Sendhoff , Trade-off between Performance and Robustness: An Evolutionary Multiobjective Approach , Second International Conference on Evolutionary Multicriterion Optimization ( Springer , 2003 ) . Google Scholar -
G. Venter , R. T. Haftka and J. Sobieszczanski-Sobieski , Robust Design Using Particle Swarm and Genetic Algorithm Optimization , 5th World Congress of Structural and Multidisciplinary Optimization ( 2003 ) . Google Scholar J. Branke , Creating Robust Solutions by Means of Evolutionary Algorithms, 1498th edn.,Lecture Notes in Computer Science (Springer, 1998) pp. 119–128. Crossref, Google Scholar-
D. Lim , Y.-S. Ong and B.-S. Lee , Inverse Multi-Objective Robust Evolutionary Design Optimization in the Presence of Uncertainty , GECCO'05 ( 2005 ) . Google Scholar -
K. Deb and H. Gupta , Evolutionary Multi-Criterion Optimization ( Springer-Verlag , 2005 ) . Crossref, Google Scholar


