Ant Colony Sample Clauses

Ant Colony. The Ant Colony optimisation algorithm is another nature-inspired approach, this time taking its source from the way ants navigate. In nature, ant colonies use pheromone trails to establish and navigate efficient routes between their nests and food sites. Although the pheromones evaporate over time, good routes are reinforced and retained as more ants pass through it, while bad routes are left to decay and eventually disappear [14]. The algorithm generally works as follows: • An initial ant colony is generated. • The ants conduct a local search and establish initial pheromone trails. • The better trails (e.g. those leading to better solutions) are reinforced by more ants over a series of iterations. • The strongest trail is followed and a new ant colony established; the process then repeats. The Ant Colony approach can have similar performance to Xxxx search [14] but can also be more computationally expensive than other approaches, e.g. penalty-based GA. As with the other algorithms, modifications and variants exist, e.g. ones which allow the use of a priori knowledge or which store and rank a number of best individuals rather than only a single individual. Using multiple individuals allows a larger portion of the design space to be explored (essentially, by multiple ant colonies) at the expense of performance.