Science and Engineering

Statistics and operations research

Overview

What is statistics and operations research?

Our discipline is comprised of researchers and educators actively engaged in the fields of:

  • data analytics
  • statistical methodology
  • probabilistic modelling
  • operations modelling
  • optimisation of systems and processes.

Collectively, we're focused on developing and examining methods for identifying useful information from noisy data, and providing optimised solutions for complex decision-making processes. We achieve this through the construction and testing of operational and stochastic models of complex systems, to better understand dynamic relationships.

Our experts in statistics and operations research work together to provide a comprehensive coverage of the three phases of analytics, providing insight and answers to three significant questions:

  • descriptive (what happened?)
  • predictive (what will happen?)
  • prescriptive (what should we do?).

This comprehensive data-analytics capability allows our partner organisations to model and understand their environment, markets and performance, and to optimise their decision-making at the strategic, tactical and operational levels.

Research

Our research focus areas and their practical applications include:

Simulation-based model validation
Assessment of external accuracy through re-sampling and cross-validation techniques
Large-scale data mining
Machine learning techniques and computationally intensive algorithms, such as random forests
Applied Bayesian statistics
Bayesian networks and hierarchical models
Bayesian computational statistics
Markov chain Monte Carlo applications to complex model fitting
Linear and integer programming
Large and complex decision problems, such as supply chain optimisation, production planning and portfolio optimisation
Meta-heuristics
Computational techniques for intractable optimisation problems
Scheduling optimisation
Scheduling and resource allocation problems, including hospital theatre and bed planning, mine planning, transport and logistics
Discrete-event simulation
Operational modelling and 'what if' analysis, for applications including rail yard operations, hospital patient flow and port operations.

Rankings

The quality of our work in the field of statistics was recognised with a rating of 4 (above world standard) from Excellence in Research for Australia (ERA).

ERA (Excellence in Research for Australia) evaluates the quality of research undertaken in Australian universities against national and international benchmarks.

Teaching

Real-world learning

Students in our capstone Operations Research unit have worked with industry partner Seqwater to optimise contingency water carting operations in South East Queensland for off-grid communities during times of drought.

This unit provides the opportunity to conduct an operations modelling analysis to address an authentic real-world problem, and provide actionable outcomes for our industry partners.

Students become industry-ready: they learn about stakeholder engagement, work in a project team within a compliance framework, and learn the language of industrial projects.

Projects

The Category 1 funded research projects we are currently leading are:

Developing novel statistical tools for more accurate predictions

This project intends to develop novel statistical tools for more accurate prediction by taking account of model complexity and uncertainties associated with the fitting procedure. The project also plans to develop a novel shrinkage approach via new penalty functions to avoid over-fitting and asymptotic properties. The key applications may include genetic studies where the number of predictors is large and biological experiments where multivariate and temporal data are often collected - for example economical breeding in animal and fish farming and more effectively detecting the genes of interest in genetic studies on human, animals and plants.
Project leader
Professor You-Gan Wang
Dates
2016-2019
Project summary
This project intends to develop novel statistical tools for more accurate prediction by taking account of model complexity and uncertainties associated with the fitting procedure. The project also plans to develop a novel shrinkage approach via new penalty functions to avoid over-fitting and asymptotic properties. The key applications may include genetic studies where the number of predictors is large and biological experiments where multivariate and temporal data are often collected - for example economical breeding in animal and fish farming and more effectively detecting the genes of interest in genetic studies on human, animals and plants.

Computationally efficient Bayesian algorithms estimating complex models

Project leader
Dr Christopher Drovandi
Dates
2016-2019
Project summary
Bayesian parameter estimation and model validation procedures are currently computationally intractable for many complex models of interest in science and technology. These included biological processes such as the efficacy of heart disease, wound healing and skin cancer treatments.
Potential outcomes of the project include new algorithms to significantly economise computations and improved understanding of the mechanisms of experiemental data generation. Imrpoved models of wound healing, skin cancer growth and heart physiology supported by these algorithms could improve population population health.

An integrated mathematical approach to synchronise and optimise hospital operations

Project leader
Adjunct Professor Erhan Kozan
Dates
2014-2017
Project summary
This project aims to develop an integrated mathematical approach to synchronise and optimise patient scheduling systems of different departments to ensure that the hospital's assets and related resources are used efficiently. The project's aim is to investigate patient flow, process delay, and the interaction and inter-dependence of departments within the hospital to reduce access block (bottleneck) and subsequent overcrowding. This project aims to smooth the running of the hospital, improve the efficiency of patient throughput, reduce waiting times, and revolutionise hospital planning and scheduling.

Bayesian statistics model to address climate change effects on the Great Barrier Reef and better understanding of neurological diseases

Project leader
Professor Kerrie Mengersen
Dates
2014-2016
Project summary
Bayesian statistics is a fundamental statistical and machine learning approach for density estimation, data analysis and inference. However, there remain open questions regarding the formulation of the model, the likelihood and priors, and efficient computation. This project proposes new approaches that address these issues, and applies them to two exemplar challenges: the impact of climate change on the Great Barrier Reef and better understanding neurological diseases related aging, in particular Parkinson's Disease.

Large-scale statistical machine learning

Project leader
Professor Peter Bartlett
Dates
2011-2016
Project summary
This research program aims to develop the science behind statistical decision problems as varied as web retrieval, genomic data analysis and financial portfolio optimisation. Advances will have a very significant practical impact in the many areas of science and technology that need to make sense of large, complex data streams.

Interdisciplinary and inter-institution projects

Some of the projects we are contributing to with other disciplines and institutions are:

Student topics

Are you looking to further your career by pursuing study at a higher and more detailed level? We are currently looking for students to research these topics:

Contact

School of Mathematical Sciences

  • Level 6, O Block, Room O617
    Gardens Point