Summary
Ant colony optimisation (ACO) is a family of search algorithms inspired by social insects. Artificial ants incrementally build candidate solutions while consulting both local heuristic information and a shared pheromone trail left by earlier successful ants. Over time, good routes become easier to rediscover while pheromone evaporation prevents the search from freezing too quickly. (Brabazon, O’Neill and McGarraghy, Natural Computing Algorithms, see source-natural-computing-algorithms)
Key ideas
- Constructive search: ants build a route step by step.
- Pheromone: stores historical evidence that a choice was successful.
- Evaporation: reduces old influence so the system can adapt.
- Heuristic desirability: local guidance such as distance or cost.
In practice
ACO is most relevant when a game problem can be treated as route or ordering construction:
- path and route optimisation
- travelling-salesman-style ordering
- network or resource-routing problems
Evidence
- The source treats ant algorithms as a distinct family of social search methods especially suited to constructive routing problems.
- Canonical ant systems combine pheromone information with heuristic desirability rather than relying on either alone.
Implications
- ACO is more naturally tied to routing and combinatorial path construction than PSO, which is usually more comfortable in continuous spaces.
- For teaching, ACO also gives a useful bridge between biological metaphor and practical optimisation.
Open questions
- Would the best classroom example be path routing, patrol-route construction, or puzzle generation?
- How should ACO be contrasted with classical deterministic pathfinding like A*?
Related
pathfinding-algorithms · particle-swarm-optimisation · genetic-algorithms · source-natural-computing-algorithms