GA4Qsys

Density functional theory (DFT)-based Genetic algorithm (GA) code for structural optimization of supported nanoparticle model in a queueing system. This code is strongly based on atomic simulation environment (ASE) python module.

Features:

• A criterion that terminates the GA if no new candidate to fill population appears for N_patience iterations was newly included.
• Issue of overwriting existing inputs for calculator (e.g., INCAR, KPOINTS, POSCAR, etc. for VASP) was resolved by generating the directory corresponding to each relaxation.

The flow of the code and data, and the crontab-based repeated interaction with a queueing system were shown in the following image.

The flowchart above was re-designed based on the figures in the original paper for the GA code implemented in the ASE module [J. Chem. Phys. 2014, 141, 044711].

How to use

link to codes

1. Prepare essential codes for the calculation in the same directory:

   calc.py, ga_init.py, ga_main_queue.py, POSCAR (to be used as support)

2. Set computational parameters (e.g., for VASP) in the calc.py file.

3. Generate initial random supported metal structures for initial population using ga_init.py code. Detailed settings for the generation can be defined in the code.

4. Run the code, ga_main_queue.py periodically using tools like crontab. Detailed parameters for the GA, such as termination criterion, maximum number of calculations in the queue, and population size, can be set in the code:

   # Parameters to be defined by a user -----------------------------------------
   directory_name       = 'Pd8_CeO2_111_GA' # for termination of crontab
   job_prefix           = 'Pd8-CeO2_111_opt' # for job names
   n_converge           = 70
   population_size      = 20
   mutation_probability = 0.3
   n_simul              = 11
   # ----------------------------------------------------------------------------
   

Application of this code

This code has been utilized in the following published papers:

1. Comparison of Pd metal dispersions on two difference CeO2 facet:
   D. Shin et al, ACS Catal. 2022, 12, 8, 4402–4414. (https://doi.org/10.1021/acscatal.2c00476)
2. Structural optimization of Pd8 nano-cluster on Co3O4 surface:
   D. Shin et al, ACS Catal. 2022, 12, 13, 8082–8093. (https://doi.org/10.1021/acscatal.2c02370)
3. Structural optimization of V_xO_y cluster on two TiO2 surfaces with difference crystallographic phases:
   D. Shin et al, ACS Catal. 2023, 13, 6, 3775–3787. (https://doi.org/10.1021/acscatal.2c05520)