Environment Dynamics

The grid is the realized world state of the model, so its dynamics matter both for what agents currently observe and for what later elections inherit from earlier ones.

Runtime Flow

Initialization starts from a preset color distribution and can then be locally patched with color_patches_steps and patch_power so the starting grid is less spatially uniform.

After that, each step follows the same timing. Mutation from election t is applied at the start of step t+1, elections run on the current election-time state, and global_color_dst is updated from the realized grid.

Implementation Reference

Initialization, scheduler timing, and global grid updates are documented in ParticipationModel. Area-level mutation logic is documented in Area.

Main Knobs

• mu in [0,1]: fraction of area cells recolored per mutation event
• election_impact_on_mutation >= 0: how strongly elected orderings shape mutation probabilities
• num_colors >= 2: color alphabet and election option count
• heterogeneity >= 0: spread of the preset initialization distribution
• color_patches_steps >= 0: number of initialization patching passes
• patch_power >= 0: patching preference for local consensus versus preset distribution