Over the past years, we have been extending and applying evolutionary computation techniques to solve problems in a wide range of domains. Below we describe some of our recent studies in the field. For more details on the ongoing research, please refer to the publications page.
Evolving collective behaviours in simulated robots
Evolutionary swarm robotics is a field that uses evolutionary computation techniques to evolve behaviours for robots in multi-robot systems. These multi-robot systems are inspired by swarms in nature (e.g., ants, bees and termites) and emphasise decentralised and self-organising behaviours. The field of evolutionary robotics has multiple common tasks and widely used benchmark activities such as navigation, obstacle avoidance, and phototaxis. We have been applying an evolutionary approach to learning behaviours that demonstrate emergent collective phototaxis in a swarm of simulated robots.
Evolved aesthetic analogies to improve artistic experience
It has been demonstrated that computational evolution can be utilised in the creation of aesthetic analogies between two artistic domains by the use of mapping expressions. When given an artistic input these mapping expressions can be used to guide the generation of content in a separate domain. For example, a piece of music can be used to create an analogous visual display.
Evolutionary approaches to generating graphs
Many problem domains such as simulation in spatial game theory or modelling the spread of information in a population involve the need to create classes of graphs that exhibit specific features. Often the question of interest relates to exploring the potential impact of certain features in a graph. In these scenarios, analysis can be facilitated by generating graphs that exhibit those features. We have been researching evolutionary approaches to generating graphs and analysing their properties. It involves considering dynamic properties of the graph during creation and their relationship with the underlying geometry.
Using evolutionary computation to detect structures in complex social networks
Being able to detect structures in complex social networks, such as cliques on Twitter graphs, is critical as it allows the ability to locate and identify links between structures and functions. By identifying various communities we can provide information as to how the network functions and the impact of topologies. There are a number of different approaches used in the detection of structures and this research focuses on the use of evolutionary algorithms to identify various communities.