The ARC Centre for Complex Systems (ACCS) is interested in employing some students to work on small projects at UQ over the 2004/2005 summer break (for about 10-12 weeks). The pay rates are about $1400 - $1600 per fortnight. We are particularly interested in later year students who are thinking about a research career or going on to study subjects related to complex systems science and engineering.

The projects:

Radial basis function neural networks with recursive bias field correction

In many important application areas such as control, pattern recognition, and signal processing, nonlinear adaptive systems such as radial basis function (RBF) networks are needed to approximate underlying nonlinear mappings through learning from examples. In this project, we consider the extension of RBF networks to model signal inhomogeneities due to acquisition equipments. Such low frequency artifcats (bias field) could be spatial or temporal correlated.  By adopting a Markov random field (MRF) model and a recursive bias field correction scheme, the bias field and true signal intensities can be estimated simultaneously. This project will therefore support ACCS projects on the application of intelligent complex networks in pattern recognition and signal processing.

This project will require a student with good skills in MATLAB and some knowledge of machine learning algorithms.  The work will start during the 2004/2005 summer break (for 12 weeks) and will be supervised by Professor Geoff McLachlan and Dr Angus Ng.

For further information contact Professor Geoff McLachlan (gjm@maths.uq.edu.au, ph: 3365 2150)

Power system transmission network service pricing using probabilistic analysis techniques

Power systems are complex, nonlinear and interactive systems involving generation, transmission and distribution systems to supply end users with electricity. The transmission system is an essential part for the overall power grid. The overall system performance depends largely on the capacity of the transmission network. Recent blackouts in North America in August 2003, as the most significant blackouts so far, are mainly because of insufficient infrastructure especially limited transmission network services. Generally, system stability across the power grid depends mainly on the transmission network. Although other parts of the power industry have been deregulated over the past decades, the transmission network remains regulated because of its role in system security and reliability. However, with the generation side deregulated into a competitive market, there is increasing pressure to improve the efficiency of transmission network services.

Proper transmission pricing in a market situation is the key element in improving the efficiency and reliability of the transmission network services. Traditionally, transmission pricing methods depend on power flow and system losses due to power transfer throughout the network. These techniques do not consider inputs from the social and economical side where the power grid is located. This project aims at exploring social and economical impacts as well as power flow information in the transmission network using probabilistic methods in order to form a framework of transmission pricing which can enhance the efficiency and reliability of the transmission network services.

The project requires the following skills: probabilistic & statistics analysis techniques; c/c++ or Matlab programming skills. Knowledge of power systems engineering, and knowledge of electricity market and transmission pricing are desirable.

Applications close 5 November. For further information contact Dr Zhao Yang Dong (email zdong@itee.uq.edu.au, tel. 3346 9052).

Radial Basis Function Neural Networks with Recursive Bias Field Correction

In many important application areas such as control, pattern recognition, and signal processing, nonlinear adaptive systems such as radial basis function (RBF) networks are needed to approximate underlying nonlinear mappings through learning from examples. In this project, we consider the extension of RBF networks to model signal inhomogeneities due to acquisition equipments. Such low frequency artifacts (bias field) could be spatial or temporal correlated. By adopting a Markov random field (MRF) model and a recursive bias field correction scheme, the bias field and the true signal intensities can be estimated simultaneously.

This project will require a student with good skills in MATLAB and some knowledge of machine learning algorithms. The work will start during the 2004/2005 summer break.

For further information contact Professor Geoff McLachlan (email: gjm@maths.uq.edu.au, tel: 3365 2150).

Porting Qu-Prolog to Windows

This project will support the ACCS project on the use of intelligent agents to model air traffic management (ATM) systems. Qu-Prolog is being used to implement intelligent agents in ATM modeling. Qu-Prolog is written in C++ and runs under Unix, Linux and MacOSX. The aim of this project is to port Qu-Prologto Windows.

This will require a student (or students) with good skills in C++ and Windows OS and some knowledge of Unix/Linux. Qu-Prolog is written in standard C++ and so the majority of the implementation should hopefully port without problems. The key areas of concern are things like timers, system calls and dynamic library loading. A Prolog background is not required for this project.

For further information contact: Dr Peter Robinson, Room 78-304, tel: 3365 3461, email: pjr@itee.uq.edu.au

Extending ATM Models

This project will support the ACCS project on the use of intelligent agents to model air traffic management (ATM) systems. We currently have some simple ATM models implemented in Qu-Prolog. The aim of this project is to extend these models by increasing the capabilities of the various agents in the models.

This will require a student (or students) with good Prolog skills.

For further information contact: Dr Peter Robinson, Room 78-304, tel: 3365 3461, email: pjr@itee.uq.edu.au

Model translation in XML

Behaviour Trees is a graphical notation for modelling system and software requirements. We are seeking support for building a rule-based translator that transforms Behaviour Trees into UML and/or CSP. All three notations (Behaviour Trees, UML and CSP) can be represented in XML. The translator will be operating on these XML representations. A possible Tool support for the Model-to-Model translation could be the Eclipse plug-in Tefkat. However, we are not restricted to this tool.

We are looking for 1 or 2 students. It is possible to work as a team. The work will also be supervised by our research staff. Closing date to apply for this project is 5th November. The application should list your experience/interests within IT and Software Engineering. 

For further information contact Dr Kirsten Winter (tel.: 3365 1625, email: kirsten@itee.uq.edu.au, Wednesday - Friday) or Dr Lars Grunske (tel.: 3365 1643, email: grunske@itee.uq.edu.au, after 31 October).

A visual framework for online biological sequence analysis

The aim of the project is to develop a framework – a set of web-enabled tools – for displaying results from advanced machine learning tools applied to biological sequence analysis. The framework will be developed using Java Servlets (based on e.g. JFreeChart) and provide easy-to-use components for a variety of biological sequence analysis tools (of which the Protein Prowler is the primary candidate).

Experience in Java and web programming, and interest in machine learning and computational biology/bioinformatics are required. The project starts as soon as a suitable candidate has been found and lasts for 10 weeks.

For further information contact Dr Mikael Boden (email: mikael@itee.uq.edu.au , tel: 3365 2035).