STORRE Collection: Electronic copies of Computing Science and Mathematics technical reports.Electronic copies of Computing Science and Mathematics technical reports.http://hdl.handle.net/1893/15532024-03-29T06:57:49Z2024-03-29T06:57:49ZMethods And Sources For Underpinning Airport Ground Movement Decision Support SystemsBrownlee, AlexanderAtkin, JasonWoodward, JohnBurke, Edmundhttp://hdl.handle.net/1893/309622021-05-07T01:49:32ZTitle: Methods And Sources For Underpinning Airport Ground Movement Decision Support Systems
Author(s): Brownlee, Alexander; Atkin, Jason; Woodward, John; Burke, Edmund
Abstract: The airport Ground Movement problem is concerned with the allo- cation of routes to aircraft for their travel along taxiways between the runway and the stands. It is important to find high quality solutions to this problem because it has a strong influence on the capacity of an airport and upon the environmental impact. The problem is particularly challenging. It has multiple objectives (such as minimising taxi time and fuel consumption). It also has considerable uncertainty, which arises from the complex operations of an airport. It is an active and topical area of research. A barrier to scientific research in this area is the lack of publicly available realistic data and benchmark problems. The reason for this is often concerned with commercial sensitivities. We have worked with airports and service providers to address this issue, by exploring several sources of freely-available data and developing algorithms for cleaning and processing the data into a more suitable form. The result is a system to generate datasets that are realistic, and that facilitate research with the potential to improve on real-world problems, without the confidentiality and commercial licensing issues usually associated with real airport data. Case studies with several international airports demonstrate the usefulness of the datasets. The algorithms have been implemented within three tools and made freely-available for researchers. A benchmark Ground Movement problem has also been made available, with results for an existing Ground Movement algorithm. It is intended that these contributions will underpin the advance of research in this difficult application area.Crowd Sourcing The Sounds Of Places With A Web Based Genetic Algorithm TechreportBrownlee, AlexanderKim, Suk-JunWang, Szu-HanChan, StellaLawson, Jamie Ahttp://hdl.handle.net/1893/293102021-05-07T01:43:13Z2019-01-01T00:00:00ZTitle: Crowd Sourcing The Sounds Of Places With A Web Based Genetic Algorithm Techreport
Author(s): Brownlee, Alexander; Kim, Suk-Jun; Wang, Szu-Han; Chan, Stella; Lawson, Jamie A
Abstract: The sounds that we associate with particular places are tightly interwoven with our memories and sense of belonging. It is assumed that such an association is a complex process, and much of its mechanism is hidden from analytical examination. The association of sound to place has been much explored and examined by artistic approaches. For example, soundscape composition, which makes great use of recorded and barely-processed sounds from places in the compositional practice, highlights the power of the association. However, it does not offer us a scientific insight into its process, particularly, the role of familiarity of sounds people hear and their association to specific places. We describe a platform designed to assist in gathering the sounds that a group of people associate with a place. A web-based evolutionary algorithm, with human-in-the-loop fitness evaluations, ranks and recombines sounds to find collections that the group rates as familiar. An experiment involving independent groups of people associated with four geographical locations shows that the process does indeed find sounds deemed familiar by participants.2019-01-01T00:00:00ZRelating Training Instances to Automatic Design of Algorithms for Bin Packing via Features (Detailed Experiments and Results)Brownlee, AlexanderWoodward, John RVeerapen, Nadarajenhttp://hdl.handle.net/1893/269572021-04-29T08:46:17Z2018-04-07T00:00:00ZTitle: Relating Training Instances to Automatic Design of Algorithms for Bin Packing via Features (Detailed Experiments and Results)
Author(s): Brownlee, Alexander; Woodward, John R; Veerapen, Nadarajen
Abstract: Automatic Design of Algorithms (ADA) shifts the burden of algorithm choice and design from developer to machine. Constructing an appropriate solver from a set of problem instances becomes a machine learning problem, with instances as training data. An efficient solver is trained for unseen problem instances with similar characteristics to those in the training set. However, this paper reveals that, as with classification and regression, for ADA not all training sets are equally valuable. We apply a typical genetic programming ADA approach for bin packing problems to several new and existing public benchmark sets. Algorithms trained on some sets are general and apply well to most others, whereas some training sets result in highly specialised algorithms that do not generalise. We relate these findings to features (simple metrics) of instances. Using instance sets with narrowly-distributed features for training results in highly specialised algorithms, whereas those with well-spread features result in very general algorithms. We show that variance in certain features has a strong correlation with the generality of the trained policies. Our results provide further grounding for recent work using features to predict algorithm performance, and show the suitability of particular instance sets for training in ADA for bin packing. The data sets, including all computed features, the evolved policies, and their performances, and the visualisations for all feature sets, are available from http://hdl.handle.net/11667/108.2018-04-07T00:00:00ZInvestigating Benchmark Correlations when Comparing Algorithms with Parameter Tuning (Detailed Experiments and Results)Christie, Lee ABrownlee, AlexanderWoodward, John Rhttp://hdl.handle.net/1893/269562021-04-29T08:46:47Z2018-04-30T00:00:00ZTitle: Investigating Benchmark Correlations when Comparing Algorithms with Parameter Tuning (Detailed Experiments and Results)
Author(s): Christie, Lee A; Brownlee, Alexander; Woodward, John R
Abstract: Benchmarks are important to demonstrate the utility of optimisation algorithms, but there is controversy about the practice of benchmarking; we could select instances that present our algorithm favourably, and dismiss those on which our algorithm under-performs. Several papers highlight the pitfalls concerned with benchmarking, some of which concern the context of the automated design of algorithms, where we use a set of problem instances (benchmarks) to train our algorithm. As with machine learning, if the training set does not reflect the test set, the algorithm will not generalize. This raises some open questions concerning the use of test instances to automatically design algorithms. We use differential evolution, and sweep the parameter settings to investigate the practice of benchmarking using the BBOB benchmarks. We make three key findings. Firstly, several benchmark functions are highly correlated. This may lead to the false conclusion that an algorithm performs well in general, when it performs poorly on a few key instances, possibly introducing unwanted bias to a resulting automatically designed algorithm. Secondly, the number of evaluations can have a large effect on the conclusion. Finally, a systematic sweep of the parameters shows how performance varies with time across the space of algorithm configurations. The data sets, including all computed features, the evolved policies, and their performances, and the visualisations for all feature sets, are available from http://hdl.handle.net/11667/109.2018-04-30T00:00:00Z