[EA] Section
Last updated: October 12, 2026
- The
[EA]section is required when using EA or EA-vc, or whenappend_struc_ea = Trueis specified in the[option]section. - See also CrySPY > Search algorithms > EA
EA
| Name | Value | Default | Description |
|---|---|---|---|
n_pop | int | Population (see also Population Size) | |
n_crsov | int | Number of offspring created by crossover | |
n_perm | int | Number of offspring created by permutation | |
n_strain | int | Number of offspring created by strain | |
n_rand | int | Number of structures created randomly | |
n_elite | int | Number of elite individuals (see also Natural Selection) | |
fit_reverse | bool | False | If False, minimal search (see also Evaluate Fitness) |
n_fittest | int | 0 | Number of individuals that remain natural selection. If set to 0, all individuals are retained. |
slct_func | TNM, RLT | Function to select parents | |
t_size | int | 3 | Tournament size. Used only used slct_func = TNM. (see also Tournament Selection) |
a_rlt | float | 10.0 | Parameter for linear scaling. Used only with slct_func = RLT. (see also Roulette Selection) |
b_rlt | float | 1.0 | Parameter for linear scaling. Used only with slct_func = RLT. (see also Roulette Selection) |
crs_lat | equal, random | random | How to mix lattice vectors (see also crossover > 5. Swap the sliced halves) |
nat_diff_tole | int | 4 | Tolerance for difference in the number of atoms in crossover. (see also crossover > 6. Select the offspring with more atoms) |
ntimes | int | 1 | Number of times in permutation. |
sigma_st | float | 0.5 | Standard deviation for strain. |
maxcnt_ea | int | 50 | Maximum number of trials in EA. |
max_parent_attempts | int | 50 | (since version 1.5.0) Maximum number of parent-selection attempts per offspring in EA. The initial parent selection is included. |
maxgen_ea | int | 0 | Maximum generation. If set to 0, no upper limit is applied. |
emax_ea | float | None | Energy upper limit (eV/atom) for natural selection. |
emin_ea | float | None | Energy lower limit (eV/atom) for natural selection. |
EA-vc
- EA-vc requires the following additional variables on top of the standard EA.
- Note that in EA-vc,
emax_eaandemin_eaare used not in natural selection but when computing the convex hull. - See also CrySPY > Search algorithms > EA-vc
- See also CrySPY > Tutorial > EA-vc > Analysis and visualization for the convex hull plot.
| Name | Value | Default | Description |
|---|---|---|---|
n_add | int | Number of offspring created by addition. | |
add_max | int | 5 | (since version 1.4.1) Maximum number of atoms to add in addition |
n_elim | int | Number of offspring created by elimination. | |
elim_max | int | 5 | (since version 1.4.1) Maximum number of atoms to eliminate in elimination |
n_subs | int | Number of offspring created by substitution. | |
subs_max | int | 3 | (since version 1.4.1) Maximum number of atoms to substitute in substitution |
ref_energies | float [float …] | (since version 1.5.0) Reference energies for formation energy calculation [eV/atom]. Specify one value per atom type in the same order as atype, separated by spaces without parentheses or commas. The number of values must match the number of atom types in atype. | |
end_point | float [float …] | (deprecated since version 1.5.0) Kept for backward compatibility. Use ref_energies instead. Uses the same units and input format as ref_energies. | |
emax_ea | float | None | Energy upper limit (eV/atom) for computing the convex hull. |
emin_ea | float | None | Energy lower limit (eV/atom) for computing the convex hull. |
- The following variables have been moved to the [visual] section since version 1.5.0.
cgenwas deprecated in version 1.5.0.- The default values below apply to versions 1.4.3 and earlier. For version 1.5.0 and later, see the [visual] section.
| Name | Value | Default | Description |
|---|---|---|---|
cgen | int | None | (in versions 1.4.1–1.4.3) Which generation’s data to plot up to. If None, data will be plotted up to the latest generation. |
show_max | float | 0.2 | When plotting the convex hull, the maximum value of the y-axis (for binary systems) or the maximum hull distance (for ternary systems) is set by show_max. |
label_stable | bool | True | Whether to show stable compositions when plotting the convex hull. |
vmax | float | 0.2 | Maximum value of the colorbar representing hull distance. |
bottom_margin | float | 0.02 | Bottom margin of the y-axis for binary convex hull plot. |
fig_format | str | svg | Figure format for convex hull plot: svg, png, or pdf. |