Global crystal structure optimization with ASE (and with a little help from DFTB)

I will mainly talk about the genetic algorithm functionality in ASE, where I have implemented various tools for global optimization of bulk compounds. These comprise random structure generators, genetic operators, and a structure comparator. This has been part of an effort to develop an efficient structure prediction approach in conjunction with automatic density functional tight-binding (DFTB) parametrization strategies [1,2,3]. Related changes in the DFTB+ and QuantumEspresso calculators will also be briefly discussed. Ideas for future work include (1) expanding the set of examples/tutorials and (2) merging the 3D-periodic and non-periodic tools into one set that can handle any periodicity.

References:

1. M. Van den Bossche, H. Grönbeck and B. Hammer, J. Chem. Theory Comput. 14, 52797-2807 (2018). doi:10.1021/acs.jctc.8b00039
2. M. Van den Bossche, J. Phys. Chem. A 123, 13, 3038-3045 (2019). doi:10.1021/acs.jpca.9b00927
3. M. Van den Bossche, C. Noguera, J. Goniakowski, In Manuscript (2019).