Ants Simulation — Ant Colony Laboratory

Small decisions, shared paths

Open Settings, choose the starting conditions and press Start. Each new visit uses a random seed for three scattered food patches; the nest starts with 50 workers by default. Each explorer senses its immediate surroundings. A returning ant leaves a food trail; other ants can follow and reinforce it. Watch scattered searches become a busy route.

There is no score target. The population changes as brood matures and workers die of old age or hunger. Deliveries now fill the nest stores: abundance can support growth, while a shortage can interrupt breeding and lead to collapse.

Try your first experiment

1. Let the ants explore and enable Pheromones. Blue marks the home trail; pink marks the food trail.
2. Wait for a busy route, then use Wall to interrupt it. Leave room to go around the ends.
3. Watch how the ants search for another passage. An enclosed food patch stays inaccessible until you open a gap.
4. Erase a food patch or let it run out. Its old trail gradually fades. New patches appear naturally as the seasons change.

Controls

The simulation starts paused with Settings centered on the playfield. Choose your parameters and press Start; opening Settings later pauses the colony while you make adjustments. Closing it restores the previous running state. The playfield fills the canvas without side margins. The compact top toolbar uses icons and groups playback, painting and pheromone visibility. Adjust brush size and toggle Follow ant directly in the toolbar. Open Settings for colony parameters, seed, reset and save/load. On narrow screens, swipe the single indicator strip sideways to see every value. Use the fullscreen button on the surrounding game page.

• Food / Wall / Erase: click, tap or drag to paint. Erase removes walls, food and pheromones. The nest and its immediate surroundings are protected.
• Select: click or tap near an ant to inspect whether it is exploring or carrying food. Enable Follow ant for a closer view; Whole colony returns to the overview.
• Initial workers: choose 50–1,000 workers for the next reset. This does not add or remove living workers. The colony can fall below 50 or become extinct; workers and brood together are capped at 1,000.
• Seasonal food growth: 0× stops new patches; 1× is the default and 2× doubles their appearance rate. Existing food still spoils slowly. You can always add food with the brush.
• Exploration: adjust variation in local steering. Evaporation: adjust how quickly both trails fade.
• Pause / One step: inspect a moment and advance one simulation tick. Speed: drag the slider from 0.5× to 4× to change the pace of the entire simulation.
• Seed / Reset: restart a reproducible landscape with the chosen initial worker count and behavior settings. The dice button chooses a new seed; reusing a seed reproduces the same initial food layout even with a different worker count. Reset removes landscape edits and deliveries and waits for Start.
• Save / Load: manually store and restore the world, ants, trails and controls in this browser. There is one save slot; storage must be available. Nothing continues running while the tab is hidden. Previous development saves can be loaded; new saves use a separate slot so the previous save remains available.
• Keyboard: focus the canvas, move the cursor with arrow keys (Shift for larger moves), and press Enter to apply the selected tool or select a nearby ant. Space pauses or resumes.

Life through the seasons

A year lasts six simulation minutes, divided into 90-second spring, summer, autumn and winter periods. Food patches appear at random positions outside the nest area. Their rate changes smoothly: spring and summer are richer, autumn and winter leaner. This is a compressed seasonal model, not a regional weather forecast. The soil tint changes with the cycle.

The nest starts small and gradually expands as its worker population reaches new highs. Its outline varies with the seed; growth is visual and keeps entrances and food storage capacity unchanged. It retains its expanded size if the population falls.

The nest begins with 180 food units and holds up to 600. The green ring around it shows how full the stores are; excess deliveries are lost. The delivered counter records all successful trips, including food later eaten. Ground food gradually spoils, with greater losses during lean periods.

Workers consume nutrition as time passes. Hungry workers return to eat one stored unit; a worker carrying food can consume its load in an emergency. A fully nourished worker has about 140 simulation seconds before starvation without another meal. Lifespans vary from 480 to 840 simulation seconds, and the starting workers have different ages. Select one to inspect its age and nutrition; very hungry workers appear lighter.

When stores contain at least 25 units, the nest can invest two units per brood item. Brood takes 30 simulation seconds to mature with food available. An empty nest pauses development; 30 seconds of accumulated starvation kills that brood. Raising brood slows in autumn and winter. Worker deaths and brood losses are counted separately. Eggs and a queen are abstracted into this nursery rule, rather than simulated as individual agents.

A dead worker drops any carried unit onto the ground. Neither seasonal food nor the food brush revives an extinct colony; use Reset to begin again. Try setting food growth to zero, compare different starting populations, or help a colony through winter with a food patch.

A simplified model of collective behavior

Ants read local sensors rather than a map of all food locations. Returning ants ignore food trails and use the home gradient and their direction toward the nest. Near a clear nest entrance they prioritize docking and unloading. Ants use a light sideways separation from a few immediate neighbours. Contact adds gentle braking and a subtle turn; disable Ant collisions in Settings if you prefer. Workers brake or take a short sideways step when another body occupies their path. With collisions enabled, moving workers maintain a minimum body spacing; workers inside the nest share its interior. Workers inside the nest are hidden and leave gradually through the same entrance used on their last return, waiting if it is occupied. Workers that have not yet returned choose an initial exit. Explorers leave a home trail that is stronger near the nest. Both trails have limited intensity and evaporate with simulation time, including at accelerated speeds.

This is an experiment in emergent paths, not a biological reconstruction of a particular ant species. Complex mazes can trap or delay ants; there is no global route planner guaranteeing a solution. Excavation, an individual queen, predators and competing colonies are outside this version. One colony keeps the relationship between food, age and population visible.

For another kind of collective behavior, explore the artificial ecosystem in BioSim.

News & Release Notes

October 21, 2026 — Ants Simulation launches

Explore a living colony with births, ageing, hunger, seasonal food and a nest that gradually grows with the population. Paint food and walls, interrupt pheromone trails and watch workers find new paths. Adjust brush size, follow an ant and control speed from the toolbar. Optional collisions keep workers apart and give returning ants priority at nest entrances. Choose reproducible starting conditions and save your experiment locally.