Identification and Optimization of AB₂ Phases Using Principal Component Analysis, Evolutionary Neural Nets, and Multiobjective Genetic Algorithms

Abstract

Available data for a large number of AB₂ compounds were subjected to a rigorous study using a combination of Principal Component Analysis (PCA) technique, multiobjective genetic algorithms, and neural networks that evolved through genetic algorithms. The identification of various phases and phase-groups were very successfully done using a decision tree approach. Since the variable hyperspaces for the different phases were highly intersecting in nature, a cumulative probability index was defined for the formation of individual compounds, which was maximized along with Pauling's electronegativity difference. The resulting Pareto-frontiers provided further insight into the nature of bonding prevailing in these compounds.

Keywords:

• AB₂ compounds
• Data mining
• Decision tree
• Evolutionary algorithm
• Genetic algorithms
• Laves phase
• Multiobjective optimization
• Neural network
• Principal component analysis

ACKNOWLEDGMENT

The authors gratefully acknowledge support from the: National Science Foundation-International Materials Institute Program for the Combinatorial Sciences and Materials Informatics Collaboratory (CoSMIC-IMI), grant # DMR-08-33853 (AS, CSK, SI, KR, and NC); Air Force Office of Scientific Research, grant # FA95500610501 (CSK and KR); and Defense Advanced Research Program Agency—Center for Interfacial Engineering for MEMS, grant # HR 0011-06-1-0049 (KR). Financial support from the Academy of Finland is gratefully acknowledged by AA and NC.