Michael Kirley

A Recursive Decomposition Method for Large Scale Continuous Optimization

Publisher: Institute of Electrical and Electronics Engineers (IEEE)

Date: 10-2018

DOI: 10.1109/TEVC.2017.2778089

An Instance Space Analysis of Constrained Multiobjective Optimization Problems

Publisher: Institute of Electrical and Electronics Engineers (IEEE)

Date: 10-2023

DOI: 10.1109/TEVC.2022.3208595

An Iterative Machine Learning Approach to Informative Performance Reporting in Dynamic Multi-Objective Optimization

Publisher: ACM

Date: 15-07-2023

DOI: 10.1145/3583133.3590624

Sampling Effects on Algorithm Selection for Continuous Black-Box Optimization

Publisher: MDPI AG

Date: 11-01-2021

DOI: 10.3390/A14010019

Abstract: In this paper, we investigate how systemic errors due to random s ling impact on automated algorithm selection for bound-constrained, single-objective, continuous black-box optimization. We construct a machine learning-based algorithm selector, which uses exploratory landscape analysis features as inputs. We test the accuracy of the recommendations experimentally using res ling techniques and the hold-one-instance-out and hold-one-problem-out validation methods. The results demonstrate that the selector remains accurate even with s ling noise, although not without trade-offs.

Study of Planetary Boundary Layer, Air Pollution, Air Quality Models and Aerosol Transport Using Ceilometers in New South Wales (NSW), Australia

Publisher: MDPI AG

Date: 21-01-2022

DOI: 10.3390/ATMOS13020176

Abstract: The planetary boundary layer height (PBLH) is one of the key factors in influencing the dispersion of the air pollutants in the troposphere and, hence, the air pollutant concentration on ground level. For this reason, accurate air pollutant concentration depends on the performance of PBLH prediction. Recently, ceilometers, a lidar instrument to measure cloud base height, have been used by atmospheric scientists and air pollution control authorities to determine the mixing level height (MLH) in improving forecasting and understanding the evolution of aerosol layers above ground at a site. In this study, ceilometer data at an urban (Lidcombe) and a rural (Merriwa) location in New South Wales, Australia, were used to investigate the relationship of air pollutant surface concentrations and surface meteorological variables with MLH, to validate the PBLH prediction from two air quality models (CCAM-CTM and WRF-CMAQ), as well as to understand the aerosol transport from sources to the receptor point at Merriwa for the three case studies where high PM10 concentration was detected in each of the three days. The results showed that surface ozone and temperature had a positive correlation with MLH, while relative humidity had negative correlation. For other pollutants (PM10, PM2.5, NO2), no clear results were obtained, and the correlation depended on the site and regional emission characteristics. The results also showed that the PBLH prediction by the two air quality models corresponded reasonably well with the observed ceilometer data and the cause and source of high PM10 concentration at Merriwa can be found by using ceilometer MLH data to corroborate back trajectory analysis of the transport of aerosols to the receptor point at Merriwa. Of the three case studies, one had aerosol sources from the north and north west of Merriwa in remote NSW, where windblown dust is the main source, and the other two had sources from the south and south east of Merriwa, where anthropogenic sources dominate.

A Networked <inline-formula> <tex-math notation=LaTeX>${N}$ </tex-math> </inline-formula>-Player Trust Game and Its Evolutionary Dynam

Publisher: Institute of Electrical and Electronics Engineers (IEEE)

Date: 12-2018

DOI: 10.1109/TEVC.2017.2769081

University Of Melbourne

Location: Australia

View Organisation

Discovery Projects - Grant ID: DP160102231

Start Date: 02-2016

End Date: 12-2019

Amount: $300,000.00

Funder: Australian Research Council

Discovery Projects - Grant ID: DP180101215

Start Date: 06-2018

End Date: 12-2021

Amount: $351,251.00

Funder: Australian Research Council

Discovery Projects - Grant ID: DP0664674

Start Date: 2006

End Date: 03-2009

Amount: $125,027.00

Funder: Australian Research Council

Discovery Projects - Grant ID: DP130100845

Start Date: 2013

End Date: 12-2017

Amount: $414,444.00

Funder: Australian Research Council

Industrial Transformation Research Hubs - Grant ID: IH210100051

Start Date: 07-2022

End Date: 07-2027

Amount: $5,000,000.00

Funder: Australian Research Council

Industrial Transformation Training Centres - Grant ID: IC200100009

Start Date: 09-2021

End Date: 09-2027

Amount: $4,861,236.00

Funder: Australian Research Council