bevy_stigmergy

Fading marks on a grid that a crowd coordinates through, like ant trails. Agents deposit into shared channels that evaporate and diffuse between floor cells, never through a wall, and read them back by value or by gradient, so a crowd converges on a trail, scatters from a threat or rallies toward a moving target without any agent planning a route. The field is computed once per tick and shared by every agent.

Add it

cargo add bevy_stigmergy

The README has a quick start that is compiled on every cargo test. The crate depends only on bevy_math; it owns no resources, registers no systems and has no schedule, so it ticks wherever your game’s loop says.

What it gives you

• StigGrid<N>: N scalar channels with deposit, evaporate_diffuse, sample_cell, gradient_cell and hotspot_cell (the field).
• RallyGrid: a vector per cell, for recruiting toward a target that moves.
• CellFrame: an optional world-to-cell mapping for an axis-aligned grid.
• evaporate_diffuse_with: the same tick split into chunks for your own thread pool, bit-identical to the inline one (ticking).
• serde and reflect features on the settings types.

Use it in the Cauldron

The Cauldron takes this crate behind its bevy_stigmergy feature, for the crowd preview (issue #107): it walks N people from one area of a map to another with bevy_orca’s avoidance, and each lays a trail and a congestion mark into a StigGrid laid over the map with a CellFrame. The result is a measurement, not an edit: the two fields drawn as an overlay, how many walkers arrived, and where congestion peaks, so you can see where a gap between buildings jams before a game ships it. The settings it uses are the crate’s own ChannelDefs, which are Serialize/Deserialize behind the serde feature.

Links

• Repository · docs.rs · crates.io · CHANGELOG