Projects
- Ant Colony Optimisation applied to Multi-objective Optimisation
Daniel Angus, PhD Candidate
Ant inspired algorithms have recently gained popularity for use in
multi-objective problem domains. One specific algorithm,
Population-based ACO, which uses a population as well as the
traditional pheromone matrix, has been shown to be effective at solving
combinatorial multi-objective optimisation problems. This research aims
at extending the Population-based ACO algorithm with a crowding
population replacement scheme to increase the search efficacy and
efficiency. Several Multi-objective Travelling Salesman Problems of
varying complexity are used to determine algorithm performance. This
work was published at the IEEE 2007 Symposium on Multi Criteria
Decision Making and was fortunate to receive the award for best student
paper. A copy of the paper can be found here. - IIDLE: Immunological Inspired Distributed Learning Environment
Jason Brownlee, PhD Candidate
IIDLE is based in a holistic abstraction of the acquired immune system as a spatially distributed, circulating and heterogeneous population of specialised discrete units that provide a homogeneous defence against external pathogenic material. It is these discrete, atomic, immutable and disposable units or information packets that represent the knowledge captured by the system, and are the substrate for the maintenance and learning processes. - Optimization Algorithm Toolkit (OAT)
Jason Brownlee
The Optimization Algorithm Toolkit (OAT) project is available on SourceForge. A workbench and toolkit for developing, evaluating, and playing with classical and state-of-the-art optimization algorithms on standard benchmark problem domains; including reference algorithm implementations (Java), graphing, visualizations and much more. Algorithms include evolutionary, swarm, immune system and others. Domains include continuous function, travelling salesman, protein folding and others. - Applying Extremal Optimisation To Dynamic Optimisation Problems
Irene Moser, PhD Candidate
Extremal Optimisation is a recently conceived addition to the range of stochastic solvers. Due to its deliberate lack of convergence behaviour, it can be expected to solve dynamic problems without having to be informed when a change occurs. Moreover, the severity of change does not seriously affect algorithm performance, allowing for unpredictable fluctuations without affecting the outcome. This project examines the algorithm's behaviour on several different problem types as well as the interaction of solver mechanisms with problem intricacies, a phenomenon shared by many algorithm implementations.
- Whole of Life Engineering (Funding: AutoCRC; Partners: GKN Aerospace Engineering Services Australia, CSIRO
James Montgomery, Tim Hendtlass
The process and products of aerospace engineering activities have a long lifespan, often longer than the involvement of particular engineers or maintenance workers. During design, production and maintenance of an aircraft a large amount of knowledge is accumulated, distributed across millions of documents. Our involvement in the Whole of Life Engineering project is in the development of a semantic search engine tailored to suit the aerospace engineering industry. The engine can be used to support immediate queries of engineers and maintenance staff as well as more complex decision support systems that must extract from the stored mass of knowledge useful facts and conclusions.
|