TY - GEN
T1 - An adaptive multi-objective evolutionary algorithm with human-like reasoning for enhanced decision-making in building design
AU - Bittermann, Michael S.
AU - Sariyildiz, I. Sevil
PY - 2011
Y1 - 2011
N2 - An adaptive multi-objective genetic algorithm is presented, where a fuzzy system is used for the fitness evaluation. The adaptivity of the evolutionary algorithm refers to modifying in a measured way the degree of relaxation of the conventional Pareto dominance concept that is used to grade solutions in multi-objective space. The aim of the adaptive relaxation is to retain adequate selection pressure during the search process. The fuzzy system models human-like reasoning that is used to evaluate the suitability of candidate solutions. This way vagueness and imprecision inherent to criteria is taken care of. Next to that, due to the use of fuzzy information processing, the resulting Pareto optimal solutions may be distinguished regarding their suitability for the ultimate goal, although from the Pareto dominance viewpoint the solutions are equivalent. This yields relevant information for a decision maker, so that some of the difficulties to select among the Pareto optimal solutions are alleviated. The algorithm is implemented for a decision making problem from the domain of architecture, where an optimal spatial arrangement of a multi-functional building is sought that satisfies three soft objectives.
AB - An adaptive multi-objective genetic algorithm is presented, where a fuzzy system is used for the fitness evaluation. The adaptivity of the evolutionary algorithm refers to modifying in a measured way the degree of relaxation of the conventional Pareto dominance concept that is used to grade solutions in multi-objective space. The aim of the adaptive relaxation is to retain adequate selection pressure during the search process. The fuzzy system models human-like reasoning that is used to evaluate the suitability of candidate solutions. This way vagueness and imprecision inherent to criteria is taken care of. Next to that, due to the use of fuzzy information processing, the resulting Pareto optimal solutions may be distinguished regarding their suitability for the ultimate goal, although from the Pareto dominance viewpoint the solutions are equivalent. This yields relevant information for a decision maker, so that some of the difficulties to select among the Pareto optimal solutions are alleviated. The algorithm is implemented for a decision making problem from the domain of architecture, where an optimal spatial arrangement of a multi-functional building is sought that satisfies three soft objectives.
KW - cognitive systems
KW - evolutionary multi-objective optimization
KW - fuzzy information processing
KW - genetic algorithm
KW - Pareto dominance
UR - http://www.scopus.com/inward/record.url?scp=79961153597&partnerID=8YFLogxK
U2 - 10.1109/SMDCM.2011.5949280
DO - 10.1109/SMDCM.2011.5949280
M3 - Conference contribution
AN - SCOPUS:79961153597
SN - 9781612840697
T3 - IEEE SSCI 2011 - Symposium Series on Computational Intelligence - MCDM 2011: 2011 IEEE Symposium on Computational Intelligence in Multicriteria Decision-Making
SP - 105
EP - 112
BT - IEEE SSCI 2011 - Symposium Series on Computational Intelligence - MCDM 2011
T2 - Symposium Series on Computational Intelligence, IEEE SSCI 2011 - 2011 IEEE Symposium on Computational Intelligence in Multicriteria Decision-Making, MCDM 2011
Y2 - 11 April 2011 through 15 April 2011
ER -