Back to the program.
Niching for Single and Multiple Objective Ant Colony Optimisation
Daniel Angus
d.angus@uq.edu.au

Abstract

The Population-based Ant Colony Optimisation (PACO) algorithm is a well-known ant-inspired algorithm which, unlike traditional ACO algorithms, maintains a finite population of solutions in addition to using a traditional pheromone information structure. PACO has been demonstrated to be an efficient and effective optimisation algorithm when applied to a range of difficult single-objective, multi-objective and dynamic problem instances. By virtue of using a population structure PACO is able to be hybridised with niching techniques typically used in Genetic Algorithms such as crowding and fitness sharing. These niching techniques allow the PACO algorithm to exploit multiple regions of the search space in parallel during a single algorithm run, particularly useful for multiple objective applications where many diverse solutions across the non-dominated front are required.

Keywords

Ant Colony Optimisation