GENETIC ALGORITHM IMPLEMENTATION FOR COMBINATORIAL PROBLEMS IN VACANCY DEFECT AND SUBSTITUTIONAL DOPING ON FE3O4

• Main Author: Ijlal Wafi, Alif
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/87965
• Institution: Institut Teknologi Bandung

In this work, we propose the usage of genetic algorithm for problems regarding finding optimal configurations of vacancies and substitution-based doping in material design. The current computational power renders calculating exhaustively all possible configurations infeasible, while lack of open-source projects to alleviate this problem becomes our main motivation for the work. We implement genetic algorithm based on Python interfaced to Quantum Espressso with SCF routine as our evaluation function, with multiple chromosome schemes to handle different types of dopants targeting different sites. Our results show that genetic algorithm could find configurations with minimum energy on the search space, reducing computation costs starting from 40% from the whole search space. Furthermore, sets of configurations generated could also be used as candidates for further attempts to find globally optimal configurations, such as structure or variable cell optimization routines