Integrate
Many academic fields contribute to optimisation methodologies and technologies, but the siloed nature of these efforts creates missed opportunities. OPTIMA integrates multidisciplinary optimisation approaches, led by world‑leading investigators who span relevant disciplines, into a common and accessible toolkit. By tackling industry problems that need insight from multiple fields, OPTIMA dissolves silos and makes state‑of‑the‑art optimisation methodologies easier for industry to use.
Advance
Industrial optimisation problems are typically messier and more complex than classical textbook examples. They are large in scale, involve multiple competing objectives, contain uncertainties from many sources and require decisions across different time scales. OPTIMA advances an industry‑ready optimisation toolkit that supports these complexities, informed by our collection of real challenge problems from industry partners.
Uptake
Successful application of optimisation methods in industrial decision‑making often requires significant expertise, which is not always accessible to domain specialists who understand industry needs. Current end‑user optimisation tools offer limited feedback and guidance, which affects trust and reduces uptake. OPTIMA transforms optimisation technology by prioritising ease of use, clarity and user confidence, making advanced methods more accessible to industry.
Our Projects
We have twelve current projects with our industry partners.
OPTIMA Small Grant Scheme
The OPTIMA Small Grant Scheme was designed to further the research potential of OPTIMA members. Each applicant could apply for up to $20,000 to complete an exciting project using optimisation methodolgies.
Publications
2026
-
G. Tack et al., MiniZinc. (Apr. 30, 2026). Zenodo. doi: 10.5281/ZENODO.19906053.
Other View at DOI ↗
MiniZinc is a free and open-source constraint modelling language. You can use MiniZinc to model constraint satisfaction and optimization problems in a high-level, solver-independent way, taking advantage of a large library of pre-defined constraints. Your model is then compiled into FlatZinc, a solver input language that is understood by a wide range of solvers. -
M. A. Muñoz, L. Zhang, H. Alipour, H. A. Khorshidi, and H. Wang, ‘Special issue on optimization in practice: selected papers from OPTIMA-CON 2024,’ Optimization Letters, Apr. 2026, doi: 10.1007/s11590-026-02299-5.
Journal Article View at DOI ↗
-
G. Tack et al., MiniZinc. (Apr. 24, 2026). Zenodo. doi: 10.5281/ZENODO.19726160.
Other View at DOI ↗
MiniZinc is a free and open-source constraint modelling language. You can use MiniZinc to model constraint satisfaction and optimization problems in a high-level, solver-independent way, taking advantage of a large library of pre-defined constraints. Your model is then compiled into FlatZinc, a solver input language that is understood by a wide range of solvers. -
G. Tack et al., MiniZinc. (Jan. 23, 2026). Zenodo. doi: 10.5281/ZENODO.18348542.
Other View at DOI ↗
MiniZinc is a free and open-source constraint modelling language. You can use MiniZinc to model constraint satisfaction and optimization problems in a high-level, solver-independent way, taking advantage of a large library of pre-defined constraints. Your model is then compiled into FlatZinc, a solver input language that is understood by a wide range of solvers.
2024
-
B. Moradi, M. Kirley, and M. A. Muñoz Acosta, ‘Sensitivity Analysis of Surrogate-assisted Bilevel Optimisation,’ Proceedings of the Genetic and Evolutionary Computation Conference Companion, pp. 411–414, Jul. 2024, doi: 10.1145/3638530.3654228.
Journal Article View at DOI ↗
Bilevel optimisation problems consist of two interactive optimisation tasks, where an upper-level task must be solved subject to the optimality of a lower-level task. Recently, a range of surrogateassisted bilevel evolutionary algorithms have been proposed, where various approximation models have been used to substitute each of the upper- or lower-level objective functions, and in some cases, the mappings between the levels. These algorithms are highly engineered systems where alternative surrogate models with varying parameters are used, often combined with other mechanisms such as local search and/or knowledge sharing. Consequently, it is difficult to isolate the effect that the surrogate model has on performance. In this paper, we address this issue. Starting from a nested optimisation algorithm - a bilevel Covariance Matrix Adaptation Evolutionary Strategy as a baseline - we systematically introduce surrogate models at the upper-level, lower-level, and both levels simultaneously, as well as for the mapping between the levels. Using a suite of benchmark problems, we scrutinise algorithm performance. The results show the acute sensitivity of performance to the objective function or mapping being modelled within the hierarchical structure. We note that, in most test problems, smaller computation costs are evident when modelling the lower-level objective function.
2023
-
H. Alsouly, M. Kirley, and M. A. Muñoz, ‘Dynamic Landscape Analysis for Constrained Multiobjective Optimization Problems,’ AI 2023: Advances in Artificial Intelligence, pp. 429–441, Nov. 2023, doi: 10.1007/978-981-99-8388-9_35.
Book Chapter View at DOI ↗
-
E. Albert, M. G. de la Banda, M. Gómez-Zamalloa, M. Isabel, and P. Stuckey, ‘Optimal dynamic partial order reduction with context-sensitive independence and observers,’ Journal of Systems and Software, vol. 202, p. 111730, Aug. 2023, doi: 10.1016/j.jss.2023.111730.
Journal Article View at DOI ↗
Dynamic Partial Order Reduction (DPOR) algorithms are used in stateless model checking of concurrent programs to avoid the exploration of equivalent execution sequences. In order to detect equivalence, DPOR relies on the notion of independence between execution steps. As this notion must be approximated, it can lose precision and thus treat execution steps as interfering when they are not. Our work is inspired by recent progress in the area that has introduced more accurate ways to exploit conditional notions of independence: Context-Sensitive DPOR considers two steps p and t independent in the current state if the states obtained by executing p⋅t and t⋅p are the same; Optimal DPOR with Observers makes their dependency conditional to the existence of future events that observe their operations. This article introduces a new algorithm, Optimal Context-Sensitive DPOR with Observers, that combines these two notions of conditional independence, and goes beyond them by exploiting their synergies. The implementation of our algorithm has been undertaken within the Nidhugg model checking tool. Our experimental evaluation, using benchmarks from the previous works, shows that our algorithm is able to effectively combine the benefits of both context-sensitive and observers-based independence and that it can produce exponential reductions over both of them. -
V. L. J. Somers and I. R. Manchester, ‘Minimizing the Risk of Spreading Processes via Surveillance Schedules and Sparse Control,’ IEEE Transactions on Control of Network Systems, vol. 10, no. 1, pp. 394–406, Mar. 2023, doi: 10.1109/tcns.2022.3203359.
Journal Article View at DOI ↗
In this article, we propose an optimization framework that combines surveillance schedules and sparse control to bound the risk of spreading processes, such as epidemics and wildfires. Here, risk is the product of the probability of an undetected outbreak and the impact of that outbreak. The aim is to bound or minimize the risk by resource allocation and persistent monitoring schedules. The presented framework utilizes the properties of positive systems and exponential cone programming to provide scalable algorithms for combined surveillance and intervention problems. We demonstrate via epidemic and wildfire examples how the method can incorporate practically relevant parameters and scenarios such as a vaccination strategy for epidemics and the effect of vegetation and outbreak rate on a wildfire. Furthermore, we show how the method can be integrated with algorithms for robotic path planning to generate persistent surveillance motion plans.
2022
-
X. Wang, R. J. Hyndman, F. Li, and Y. Kang, ‘Forecast combinations: An over 50-year review,’ International Journal of Forecasting, vol. 39, no. 4, pp. 1518–1547, Oct. 2023, doi: 10.1016/j.ijforecast.2022.11.005.
Journal Article View at DOI ↗
-
N. Neelofar, K. Smith-Miles, M. A. Muñoz, and A. Aleti, ‘Instance Space Analysis of Search-Based Software Testing,’ IEEE Transactions on Software Engineering, vol. 49, no. 4, pp. 2642–2660, Apr. 2023, doi: 10.1109/tse.2022.3228334.
Journal Article View at DOI ↗
Search-based software testing (SBST) is now a mature area, with numerous techniques developed to tackle the challenging task of software testing. SBST techniques have shown promising results and have been successfully applied in the industry to automatically generate test cases for large and complex software systems. Their effectiveness, however, has been shown to be problem dependent. In this paper, we revisit the problem of objective performance evaluation of SBST techniques in light of recent methodological advances – in the form of Instance Space Analysis (ISA) – enabling the strengths and weaknesses of SBST techniques to be visualised and assessed across the broadest possible space of problem instances (software classes) from common benchmark datasets. We identify features of SBST problems that explain why a particular instance is hard for an SBST technique, reveal areas of hard and easy problems in the instance space of existing benchmark datasets, and identify the strengths and weaknesses of state-of-the-art SBST techniques. In addition, we examine the diversity and quality of common benchmark datasets used in experimental evaluations. -
N. Andrés‐Thió, M. Brazil, C. Ras, and D. Thomas, ‘Network augmentation for disaster‐resilience against geographically correlated failure,’ Networks, vol. 81, no. 4, pp. 419–444, Dec. 2022, doi: 10.1002/net.22138.
Journal Article View at DOI ↗
Abstract We introduce a formal framework for the study of augmenting networks in the plane for disaster‐resilience, where a disaster is modeled by a straight‐line segment. We generalize various graph structures from classical 2‐edge‐connectivity, including minimal cuts and blocks. The key concept that we introduce is that of an ‐leaf, which builds on the fundamental “leaf‐block” concept from classical augmentation. We present a number of algorithms for constructing the above‐mentioned graph structures, including a sweep‐line algorithm that finds all edge‐cuts that can be destroyed by a single disaster. We also present an algorithm which optimally adds a single edge between a pair of ‐leaves or blocks while avoiding certain disaster regions. Finally, we present a number of heuristic schemes for solving the disaster‐resilient network augmentation problem and perform extensive experiments to demonstrate the power of the ‐leaf concept within heuristic design. -
K. Smith-Miles and M. A. Muñoz, ‘Instance Space Analysis for Algorithm Testing: Methodology and Software Tools,’ ACM Computing Surveys, vol. 55, no. 12, pp. 1–31, Mar. 2023, doi: 10.1145/3572895.
Review View at DOI ↗
Instance Space Analysis (ISA) is a recently developed methodology to (a) support objective testing of algorithms and (b) assess the diversity of test instances. Representing test instances as feature vectors, the ISA methodology extends Rice’s 1976 Algorithm Selection Problem framework to enable visualization of the entire space of possible test instances, and gain insights into how algorithm performance is affected by instance properties. Rather than reporting algorithm performance on average across a chosen set of test problems, as is standard practice, the ISA methodology offers a more nuanced understanding of the unique strengths and weaknesses of algorithms across different regions of the instance space that may otherwise be hidden on average. It also facilitates objective assessment of any bias in the chosen test instances and provides guidance about the adequacy of benchmark test suites. This article is a comprehensive tutorial on the ISA methodology that has been evolving over several years, and includes details of all algorithms and software tools that are enabling its worldwide adoption in many disciplines. A case study comparing algorithms for university timetabling is presented to illustrate the methodology and tools. -
D. Bustos-Coral and A. M. Costa, ‘Drayage routing with heterogeneous fleet, compatibility constraints, and truck load configurations,’ Transportation Research Part E: Logistics and Transportation Review, vol. 168, p. 102922, Dec. 2022, doi: 10.1016/j.tre.2022.102922.
Journal Article View at DOI ↗
-
Z. Shireen et al., ‘A machine learning enabled hybrid optimization framework for efficient coarse-graining of a model polymer,’ npj Computational Materials, vol. 8, no. 1, Nov. 2022, doi: 10.1038/s41524-022-00914-4.
Journal Article View at DOI ↗
Abstract This work presents a framework governing the development of an efficient, accurate, and transferable coarse-grained (CG) model of a polyether material. The framework combines bottom-up and top-down approaches of coarse-grained model parameters by integrating machine learning (ML) with optimization algorithms. In the bottom-up approach, bonded interactions of the CG model are optimized using deep neural networks (DNN), where atomistic bonded distributions are matched. In the top-down approach, optimization of nonbonded parameters is accomplished by reproducing the temperature-dependent experimental density. We demonstrate that developed framework addresses the thermodynamic consistency and transferability issues associated with the classical coarse-graining approaches. The efficiency and transferability of the CG model is demonstrated through accurate predictions of chain statistics, the limiting behavior of the glass transition temperature, diffusion, and stress relaxation, where none were included in the parametrization process. The accuracy of the predicted properties are evaluated in context of molecular theories and available experimental data. -
D. Zhao, H. Wang, J. Huang, and X. Lin, ‘Insurance Contract for High Renewable Energy Integration,’ 2022 IEEE International Conference on Communications, Control, and Computing Technologies for Smart Grids (SmartGridComm), pp. 271–277, Oct. 2022, doi: 10.1109/smartgridcomm52983.2022.9960994.
Journal Article View at DOI ↗
The increasing penetration of renewable energy poses significant challenges to power grid reliability. There have been increasing interests in utilizing financial tools, such as insurance, to help end-users hedge the potential risk of lost load due to renewable energy variability. With insurance, a user pays a premium fee to the utility, so that he will get compensated in case his demand is not fully satisfied. A proper insurance design needs to resolve the following two challenges: (i) users' reliability preference is private information; and (ii) the insurance design is tightly coupled with the renewable energy investment decision. To address these challenges, we adopt the contract theory to elicit users' private reliability preferences, and we study how the utility can jointly optimize the insurance contract and the planning of renewable energy. A key analytical challenge is that the joint optimization of the insurance design and the planning of renewables is non-convex. We resolve this difficulty by revealing important structural properties of the optimal solution, using the help of two benchmark problems: the no-insurance benchmark and the social-optimum benchmark. Compared with the no-insurance benchmark, we prove that the social cost and users' total energy cost are always no larger under the optimal contract. Simulation results show that the largest benefit of the insurance contract is achieved at a medium electricity-bill price together with a low type heterogeneity and a high renewable uncertainty. -
S. Kandanaarachchi and R. J. Hyndman, ‘Anomaly detection in dynamic networks’, 2022, arXiv. doi: 10.48550/ARXIV.2210.07407.
Preprint View at DOI ↗
Detecting anomalies from a series of temporal networks has many applications, including road accidents in transport networks and suspicious events in social networks. While there are many methods for network anomaly detection, statistical methods are under utilised in this space even though they have a long history and proven capability in handling temporal dependencies. In this paper, we introduce textit{oddnet}, a feature-based network anomaly detection method that uses time series methods to model temporal dependencies. We demonstrate the effectiveness of oddnet on synthetic and real-world datasets. The R package oddnet implements this algorithm. -
K. Smith-Miles and M. A. Muñoz, ‘Optimal construction of montages from mathematical functions on a spectrum of order–disorder preference,’ Journal of Mathematics and the Arts, vol. 16, no. 4, pp. 347–373, Oct. 2022, doi: 10.1080/17513472.2022.2139663.
Journal Article View at DOI ↗
We previously generated diverse mathematical functions that are difficult for optimization algorithms. Represented as 2D contour plots, each image depicts a ‘blue river’ running through an intricate landscape. This paper describes the challenge of constructing an aesthetic montage of these images. A survey revealed a spectrum of tastes, divergent in preference from order to disorder, considering the structure created by connecting these ‘blue rivers’. A new artwork, Negentropy Triptych, was created to depict this spectrum by manually swapping images from a random arrangement, guided by human eye to enhance or destroy the structure. An optimization algorithm automates the process, with the results of its efforts to emulate the artistic vision presented and discussed. The challenges faced by the algorithm, despite exploring several objective functions, highlight the difficulties of capturing the goals that a human decision-maker can easily achieve. Therefore, machine learning of these goals is a promising future direction. -
T. C. Lopes, A. S. Michels, N. Brauner, and L. Magatão, ‘Balancing-sequencing paced assembly lines: a multi-objective mixed-integer linear case study,’ International Journal of Production Research, vol. 61, no. 17, pp. 5901–5917, Sep. 2022, doi: 10.1080/00207543.2022.2118888.
Journal Article View at DOI ↗
This paper considers the optimisation of Mixed-model assembly lines with continuous paced line control. The two minimisation goals have a mixed-integer linear multi-objective dispute. The paper proposes a criterion-space method to define the Pareto front for this class of problems. The method combines Pareto fronts obtained from integer solutions and gradually refines them until the instance's global front is determined. Comparing paced to unpaced line controls can be challenging, since they can produce the same cycle time given sufficiently long line lengths or buffers. Hence, determining Pareto fronts between cycle time and line length for paced lines allows meaningful comparisons between line controls. An industrial case study shows that line length acts as continuously distributable buffers for paced lines, leading to weaker diminishing returns. This result suggests that paced lines are more efficient than unpaced ones for lower cycle time ranges. -
E. Yap, M. A. Munoz, and K. Smith-Miles, ‘Informing Multiobjective Optimization Benchmark Construction Through Instance Space Analysis,’ IEEE Transactions on Evolutionary Computation, vol. 26, no. 6, pp. 1246–1260, Dec. 2022, doi: 10.1109/tevc.2022.3205165.
Journal Article View at DOI ↗
The role of carefully constructed benchmark suites in algorithm design and testing is critical. Within the continuous multiobjective optimization domain, existing suites include the general purpose ZDT, DTLZ, and WFG suites, and more recent ones specifically designed to explore the impacts of a particular problem characteristic. However, the relationship between existing suites is not clear, and the field would benefit from a “stock-take” assessment. This article investigates the coverage of current continuous multiobjective suites using the instance space analysis (ISA) methodology. Exploratory landscape analysis is used to measure critical features of each problem suite. Thereafter, we generate a 2-D visualization of the existing problem instances by locating them in the instance space, assessing their diversity, and identifying whether there are sparse areas of value to fill with new problem instances. Our findings show that the current suites are restricted in diversity when representing the entire problem instance space. We propose and evaluate three problem construction methods: 1) problem tuning; 2) toolkit hybridization; and 3) new function injection. Problem tuning is shown to generate problems surrounding existing instances, while hybridization creates problems falling between existing suites. Furthermore, utilizing the insights afforded by ISA, we show how problem features can be identified to inform the creation of new functions which fill gaps toward the boundaries of the instance space. -
D. B. Huberman, B. J. Reich, and H. D. Bondell, ‘Correction: Nonparametric conditional density estimation in a deep learning framework for short-term forecasting,’ Environmental and Ecological Statistics, vol. 29, no. 4, pp. 913–913, Aug. 2022, doi: 10.1007/s10651-022-00543-6.
Journal Article View at DOI ↗
-
D. Rajapaksha, C. Bergmeir, and R. J. Hyndman, ‘LoMEF: A framework to produce local explanations for global model time series forecasts,’ International Journal of Forecasting, vol. 39, no. 3, pp. 1424–1447, Jul. 2023, doi: 10.1016/j.ijforecast.2022.06.006.
Journal Article View at DOI ↗
-
B. Moya, R. Moreno, S. Püschel-Løvengreen, A. M. Costa, and P. Mancarella, ‘Uncertainty representation in investment planning of low-carbon power systems,’ Electric Power Systems Research, vol. 212, p. 108470, Nov. 2022, doi: 10.1016/j.epsr.2022.108470.
Journal Article View at DOI ↗
-
A. S. Michels and A. M. Costa, ‘Mixed-integer linear programming models for the type-II resource-constrained assembly line balancing problem,’ Assembly Automation, vol. 42, no. 5, pp. 585–594, Aug. 2022, doi: 10.1108/aa-10-2021-0140.
Journal Article View at DOI ↗
Purpose Resource-constrained assembly lines are widely found in industries that manufacture complex products. In such lines, tasks may require specific resources to be processed. Therefore, decisions on which tasks and resources will be assigned to each station must be made. When the number of available stations is fixed, the problem’s main goal becomes the minimisation of cycle time (type-II version). This paper aims to explore this variant of the problem that lacks investigation in the literature. Design/methodology/approach In this paper, the authors propose mixed-integer linear programming (MILP) models to minimise cycle time in resource-constrained assembly lines, given a limited number of stations and resources. Dedicated and alternative resource types for tasks are considered in different scenarios. Findings Besides, past modelling decisions and assumptions are questioned. The authors discuss how they were leading to suboptimal solutions and offer a rectification. Practical implications The proposed models and data set fulfil more practical concerns by taking into account characteristics found in real-world assembly lines. Originality/value The proposed MILP models are applied to an existing data set, results are compared against a constraint programming model, and new optimal solutions are obtained. Moreover, a data set extension is proposed due to the simplicity of the current one and instances up to 70 tasks are optimally solved. -
C. Cheng, J.-W. Lu, R. Zhu, Z. Xiao, A. M. Costa, and R. G. Thompson, ‘An integrated multi-objective model for disaster waste clean-up systems optimization,’ Transportation Research Part E: Logistics and Transportation Review, vol. 165, p. 102867, Sep. 2022, doi: 10.1016/j.tre.2022.102867.
Journal Article View at DOI ↗
-
T. Zhang, J. Wang, H. Wang, J. Ruiyang, G. Li, and M. Zhou, ‘On the Coordination of Transmission-Distribution Grids: A Dynamic Feasible Region Method,’ IEEE Transactions on Power Systems, vol. 38, no. 2, pp. 1857–1868, Mar. 2023, doi: 10.1109/tpwrs.2022.3197556.
Journal Article View at DOI ↗
Recently, the turnover of energy services between transmission-distribution grids has continued to increase, arousing widespread attention to the efficient coordination between transmission system operator (TSO) and distribution system operator (DSO). The existing literature has characterized the feasible region of distribution networks via equivalent projection methods, which is conducive to the participation of DSO in TSO scheduling. However, the redundant constraints of DSO and the processing of temporal-coupled constraints remain challenging in realistic cases. To achieve an effective interaction between TSO and DSO with least information, we develop a TSO-DSO coordination framework with temporal-coupled constraints embedded, i.e., the concept of dynamic feasible region (DFR). We theoretically derive the polyhedral form of a DFR, which can be characterized as a constraint set formed by the extreme points of the dual space of the DSO optimization program. A novel outer progressive approximation (OPA)-based algorithm is designed to effectively add feasibility cuts until convergence is achieved. To reduce the redundant constraints in the DSO operation model, the enhanced umbrella constraint identification (E-UCI) method is employed. Case studies based on i) a real-world distribution grid in China and IEEE 30-bus system, ii) a Caracas 141-bus distribution network and IEEE 118-bus transmission grid are adopted to validate the effectiveness and computational efficiency of our proposed method. -
N. Andrés-Thió, M. A. Muñoz, and K. Smith-Miles, ‘Bifidelity Surrogate Modelling: Showcasing the Need for New Test Instances,’ INFORMS Journal on Computing, vol. 34, no. 6, pp. 3007–3022, Nov. 2022, doi: 10.1287/ijoc.2022.1217.
Journal Article View at DOI ↗
In recent years, multifidelity expensive black-box (Mf-EBB) methods have received increasing attention due to their strong applicability to industrial design problems. The challenge, however, is that knowledge of the relationship between decisions and objective values is limited to a small set of sample observations of variable quality. In the field of Mf-EBB, a problem instance consists of an expensive yet accurate source of information, and one or more cheap yet less accurate sources of information. The field aims to provide techniques either to accurately explain how decisions affect design outcome, or to find the best decisions to optimise design outcomes. Many techniques that use surrogate models have been developed to provide solutions to both aims. Only in recent years, however, have researchers begun to explore the conditions under which these new techniques are reliable, often focusing on problems with a single low-fidelity function, known as bifidelity expensive black-box (Bf-EBB) problems. This study extends the existing Bf-EBB test instances found in the literature, as well as the features used to determine when the low-fidelity information source should be used. A literature test suite is constructed and augmented with new instances to demonstrate the potentially misleading results that could be reached using only the instances currently found in the literature, and to expose the criticality of a more heterogeneous test suite for algorithm assessment. Addressing the shortcomings of the existing literature, a new set of features is presented, as well as a new instance creation procedure, and a study of their impact on algorithm assessment is conducted. The low-fidelity information source is shown to be valuable if it is often locally accurate, even when its overall accuracy is relatively low. This contradicts the existing literature guidelines, which indicate the low-fidelity information is only useful if it has a high overall accuracy. History: Accepted by Antonio Frangioni, Area Editor for Design & Analysis of Algorithms – Continuous. Funding: This work was supported by Australian Research Council [Grant IC200100009] for the ARC Training Centre in Optimisation Technologies, Integrated Methodologies and Applications (OPTIMA), and the University of Melbourne Research Computing Services and Petascale Campus Initiative. N. Andrés-Thió is also supported by a Research Training Program scholarship from the University of Melbourne. Supplemental Material: The software that supports the findings of this study is available within the paper and its Supplementary Information [ https://pubsonline.informs.org/doi/suppl/10.1287/ijoc.2022.1217 ] or is available from the IJOC GitHub software repository ( https://github.com/INFORMSJoC ) at [ http://dx.doi.org/10.5281/zenodo.6578060 ]. -
R. Moss et al., ‘Forecasting COVID-19 activity in Australia to support pandemic response: May to October 2020,’ Aug. 2022, doi: 10.1101/2022.08.04.22278391.
Preprint View at DOI ↗
Abstract As of January 2021, Australia had effectively controlled local transmission of COVID-19 despite a steady influx of imported cases and several local, but contained, outbreaks in 2020. Throughout 2020, state and territory public health responses were informed by weekly situational reports that included an ensemble forecast for each jurisdiction. We present here an analysis of one forecasting model included in this ensemble across the variety of scenarios experienced by each jurisdiction from May to October 2020. We examine how successfully the forecasts characterised future case incidence, subject to variations in data timeliness and completeness, showcase how we adapted these forecasts to support decisions of public health priority in rapidly-evolving situations, evaluate the impact of key model features on forecast skill, and demonstrate how to assess forecast skill in real-time before the ground truth is known. Conditioning the model on the most recent, but incomplete, data improved the forecast skill, emphasising the importance of developing strong quantitative models of surveillance system characteristics, such as ascertainment delay distributions. Forecast skill was highest when there were at least 10 reported cases per day, the circumstances in which authorities were most in need of forecasts to aid in planning and response. -
A. Panagiotelis, P. Gamakumara, G. Athanasopoulos, and R. J. Hyndman, ‘Probabilistic forecast reconciliation: Properties, evaluation and score optimisation,’ European Journal of Operational Research, vol. 306, no. 2, pp. 693–706, Apr. 2023, doi: 10.1016/j.ejor.2022.07.040.
Journal Article View at DOI ↗
-
J. Liu, K. Marriott, T. Dwyer, and G. Tack, ‘Increasing User Trust in Optimisation through Feedback and Interaction,’ ACM Transactions on Computer-Human Interaction, vol. 29, no. 5, pp. 1–34, Oct. 2022, doi: 10.1145/3503461.
Journal Article View at DOI ↗
User trust plays a key role in determining whether autonomous computer applications are relied upon. It will play a key role in the acceptance of emerging AI applications such as optimisation. Two important factors known to affect trust are system transparency, i.e., how well the user understands how the system works, and system performance. However, in the case of optimisation, it is difficult for the end-user to understand the underlying algorithms or to judge the quality of the solution. Through two controlled user studies, we explore whether the user is better able to calibrate their trust in the system when: (a) They are provided feedback on the system operation in the form of visualisation of intermediate solutions and their quality; (b) They can interactively explore the solution space by modifying the solution returned by the system. We found that showing intermediate solutions can lead to over-trust, while interactive exploration leads to more accurately calibrated trust. -
S. Ahmadi, G. Tack, D. Harabor, and P. Kilby, ‘Weight Constrained Path Finding with Bidirectional A*,’ Proceedings of the International Symposium on Combinatorial Search, vol. 15, no. 1, pp. 2–10, Jul. 2022, doi: 10.1609/socs.v15i1.21746.
Journal Article View at DOI ↗
Weight constrained path finding, known as a challenging variant of the classic shortest path problem, aims to plan cost optimum paths whose weight/resource usage is limited by a side constraint. Given the bi-criteria nature of the problem (i.e., the presence of cost and weight), solutions to the Weight Constrained Shortest Path Problem (WCSPP) have some properties in common with bi-objective search. This paper leverages the state-of-the-art bi-objective search algorithm BOBA* and presents WC-BA*, an exact A*-based WCSPP method that explores the search space in different objective orderings bidirectionally. We also enrich WC-BA* with two novel heuristic tuning approaches that can significantly reduce the number of node expansions in the exhaustive search of A*. The results of our experiments on a large set of realistic problem instances show that our new algorithm solves all instances and outperforms the state-of-the-art WCSPP algorithms in various scenarios. -
N. James and H. Bondell, ‘Temporal and spectral governing dynamics of Australian hydrological streamflow time series,’ Journal of Computational Science, vol. 63, p. 101767, Sep. 2022, doi: 10.1016/j.jocs.2022.101767.
Journal Article View at DOI ↗
-
M. A. Muñoz, ‘Examining algorithm behavior using recurrence quantification and landscape analyses,’ Proceedings of the Genetic and Evolutionary Computation Conference Companion, pp. 1658–1665, Jul. 2022, doi: 10.1145/3520304.3534029.
Journal Article View at DOI ↗
Differences in performance between algorithms can be attributed to the interaction between their unique rule-sets and the characteristics of the instance's landscape. However, understanding this interaction can be difficult because algorithms are often composed of multiple elements, and in the worst cases are described using opaque notation and metaphors. In this paper, we introduce a methodology for the behavioral analysis of optimization algorithms, based on comparing algorithm dynamics in a given problem instance. At the methodology's core lays the hypothesis that if two algorithms, with the exact same initial conditions, have similar dynamics, then their rule-sets are also similar. An examination of Grey Wolf Optimization, shows that it exhibits bias leading to similar behavioral patterns regardless of the function. -
M. A. Muñoz, H. Soleimani, and S. Kandanaarachchi, ‘Benchmarking algorithm portfolio construction methods,’ Proceedings of the Genetic and Evolutionary Computation Conference Companion, pp. 499–502, Jul. 2022, doi: 10.1145/3520304.3528880.
Journal Article View at DOI ↗
A portfolio is a set of algorithms, which run concurrently or interchangeably, whose aim is to improve performance by avoiding a bad selection of a single algorithm. Despite its high error tolerance, a carefully constructed portfolio, i.e., the smallest set of complementary algorithms, is expected to perform better than an arbitrarily constructed one. In this paper, we benchmark five algorithm portfolio construction methods, using as benchmark problems the ASLib scenarios, under a cross-validation regime. We examine the performance of each portfolio in terms of its riskiness, i.e., the existence of unsolved problems on the test set, and its robustness, i.e., the existence of an algorithm that solves most instances. The results demonstrate that two of these methods produce portfolios with the lowest risk, albeit with different levels of robustness. -
D. Herring, M. Kirley, and X. Yao, ‘Reproducibility and baseline reporting for dynamic multi-objective benchmark problems,’ Proceedings of the Genetic and Evolutionary Computation Conference, pp. 529–537, Jul. 2022, doi: 10.1145/3512290.3528791.
Journal Article View at DOI ↗
Dynamic multi-objective optimization problems (DMOPs) are widely accepted to be more challenging than stationary problems due to the time-dependent nature of the objective functions and/or constraints. Evaluation of purpose-built algorithms for DMOPs is often performed on narrow selections of dynamic instances with differing change magnitude and frequency or a limited selection of problems. In this paper, we focus on the reproducibility of simulation experiments for parameters of DMOPs. Our framework is based on an extension of PlatEMO, allowing for the reproduction of results and performance measurements across a range of dynamic settings and problems. A baseline schema for dynamic algorithm evaluation is introduced, which provides a mechanism to interrogate performance and optimization behaviours of well-known evolutionary algorithms that were not designed specifically for DMOPs. Importantly, by determining the maximum capability of non-dynamic multi-objective evolutionary algorithms, we can establish the minimum capability required of purpose-built dynamic algorithms to be useful. The simplest modifications to manage dynamic changes introduce diversity. Allowing non-dynamic algorithms to incorporate mutated/random solutions after change events determines the improvement possible with minor algorithm modifications. Future expansion to include current dynamic algorithms will enable reproduction of their results and verification of their abilities and performance across DMOP benchmark space. -
P. Y. A. Paiva, C. C. Moreno, K. Smith-Miles, M. G. Valeriano, and A. C. Lorena, ‘Relating instance hardness to classification performance in a dataset: a visual approach,’ Machine Learning, vol. 111, no. 8, pp. 3085–3123, Jun. 2022, doi: 10.1007/s10994-022-06205-9.
Journal Article View at DOI ↗
Machine Learning studies often involve a series of computational experiments in which the predictive performance of multiple models are compared across one or more datasets. The results obtained are usually summarized through average statistics, either in numeric tables or simple plots. Such approaches fail to reveal interesting subtleties about algorithmic performance, including which observations an algorithm may find easy or hard to classify, and also which observations within a dataset may present unique challenges. Recently, a methodology known as Instance Space Analysis was proposed for visualizing algorithm performance across different datasets. This methodology relates predictive performance to estimated instance hardness measures extracted from the datasets. However, the analysis considered an instance as being an entire classification dataset and the algorithm performance was reported for each dataset as an average error across all observations in the dataset. In this paper, we developed a more fine-grained analysis by adapting the ISA methodology. The adapted version of ISA allows the analysis of an individual classification dataset by a 2-D hardness embedding, which provides a visualization of the data according to the difficulty level of its individual observations. This allows deeper analyses of the relationships between instance hardness and predictive performance of classifiers. We also provide an open-access Python package named PyHard, which encapsulates the adapted ISA and provides an interactive visualization interface. We illustrate through case studies how our tool can provide insights about data quality and algorithm performance in the presence of challenges such as noisy and biased data. -
X. Wang, Y. Kang, R. J. Hyndman, and F. Li, ‘Distributed ARIMA models for ultra-long time series,’ International Journal of Forecasting, vol. 39, no. 3, pp. 1163–1184, Jul. 2023, doi: 10.1016/j.ijforecast.2022.05.001.
Preprint View at DOI ↗
-
V. L. J. Somers and I. R. Manchester, ‘Multi-Stage Sparse Resource Allocation for Control of Spreading Processes over Networks,’ 2022 American Control Conference (ACC), pp. 3632–3639, Jun. 2022, doi: 10.23919/acc53348.2022.9867834.
Journal Article View at DOI ↗
In this paper we propose a method for sparse dynamic allocation of resources to bound the risk of spreading processes, such as epidemics and wildfires, using convex optimization and dynamic programming techniques. Here, risk is defined as the risk of an undetected outbreak, i.e. the product of the probability of an outbreak occurring and the future impact of that outbreak, and we can allocate budgeted resources each time step to bound and minimize the risk. Our method in particular provides sparsity of resources, which is important due to the large network structures involved with spreading processes and has advantages when resources can not be distributed widely. -
I. Grossman, K. Bandara, T. Wilson, and M. Kirley, ‘Can machine learning improve small area population forecasts? A forecast combination approach,’ Computers, Environment and Urban Systems, vol. 95, p. 101806, Jul. 2022, doi: 10.1016/j.compenvurbsys.2022.101806.
Journal Article View at DOI ↗
-
S. Kandanaarachchi, H. Ochiai, and A. Rao, ‘Honeyboost: Boosting honeypot performance with data fusion and anomaly detection,’ Expert Systems with Applications, vol. 201, p. 117073, Sep. 2022, doi: 10.1016/j.eswa.2022.117073.
Journal Article View at DOI ↗
-
W. D. Xu, M. J. Burns, F. Cherqui, K. Smith‐Miles, and T. D. Fletcher, ‘Coordinated Control Can Deliver Synergies Across Multiple Rainwater Storages,’ Water Resources Research, vol. 58, no. 2, Feb. 2022, doi: 10.1029/2021wr030266.
Journal Article View at DOI ↗
Abstract Studies in Real−Time Control (RTC) Rainwater Harvesting Systems (RWH) have to date been limited to the control of single storages, leaving the potential benefits of operating multiple storages in a coordinated manner largely untested. In this study, we aimed to design an optimization‐based RTC strategy that can operate multiple storages in a coordinated manner to achieve multiple objectives. We modeled the long‐term performance of this coordinated approach (i.e., termed as coordinated control ) across a range of storage sizes and compared it with a strategy that optimized the operation of each storage individually, ignoring the state of other stores within the system. Our results show that coordinated control delivered a synergy benefit in achieving better baseflow restoration, with almost no detriment to the water supply and flood protection (overflow reduction) performance. The efficiency achieved through coordinated control allows large storages to compensate for smaller, underperforming systems, to achieve higher overall performance. Such a finding suggests a general control principle in building coordination among multiple storages, which can potentially be adapted to mitigate flooding risks, and also applied to other stormwater control measures. This also opens up a new opportunity for practitioners to construct a future “smart rainwater grid” using a network of distributed storages, in combination with centralized large storages, to manage urban stormwater in a range of contexts and for a range of environmental objectives. -
N. James, M. Menzies, and H. Bondell, ‘In search of peak human athletic potential: A mathematical investigation,’ Chaos: An Interdisciplinary Journal of Nonlinear Science, vol. 32, no. 2, Feb. 2022, doi: 10.1063/5.0073141.
Journal Article View at DOI ↗
This paper applies existing and new approaches to study trends in the performance of elite athletes over time. We study both track and field scores of men and women athletes on a yearly basis from 2001 to 2019, revealing several trends and findings. First, we perform a detailed regression study to reveal the existence of an "Olympic effect," where average performance improves during Olympic years. Next, we study the rate of change in athlete performance and fail to reject the notion that athlete scores are leveling off, at least among the top 100 annual scores. Third, we examine the relationship in performance trends among men and women's categories of the same event, revealing striking similarity, together with some anomalous events. Finally, we analyze the geographic composition of the world's top athletes, attempting to understand how the diversity by country and continent varies over time across events. We challenge a widely held conception of athletics that certain events are more geographically dominated than others. Our methods and findings could be applied more generally to identify evolutionary dynamics in group performance and highlight spatiotemporal trends in group composition. -
P. Sritharan, M. A. Muñoz, P. Pivonka, A. L. Bryant, H. Mokhtarzadeh, and L. G. Perraton, ‘Biomechanical Markers of Forward Hop-Landing After ACL-Reconstruction: A Pattern Recognition Approach,’ Annals of Biomedical Engineering, vol. 50, no. 3, pp. 330–342, Jan. 2022, doi: 10.1007/s10439-022-02921-4.
Journal Article View at DOI ↗
Biomechanical changes after anterior cruciate ligament reconstruction (ACLR) may be detrimental to long-term knee-joint health. We used pattern recognition to characterise biomechanical differences during the landing phase of a single-leg forward hop after ACLR. Experimental data from 66 individuals 12-24 months post-ACLR (28.2 ± 6.3 years) and 32 controls (25.2 ± 4.8 years old) were input into a musculoskeletal modelling pipeline to calculate joint angles, joint moments and muscle forces. These waveforms were transformed into principal components (features), and input into a pattern recognition pipeline, which found 10 main distinguishing features (and 8 associated features) between ACLR and control landing biomechanics at significance [Formula: see text]. Our process identified known biomechanical characteristics post-ACLR: smaller knee flexion angle; less knee extensor moment; lower vasti, rectus femoris and hamstrings forces. Importantly, we found more novel and less well-understood adaptations: smaller ankle plantar flexor moment; lower soleus forces; and altered patterns of knee rotation angle, hip rotator moment and knee abduction moment. Crucially, we identified, with high certainty, subtle aberrations indicating landing instability in the ACLR group for: knee flexion and internal rotation angles and moments; hip rotation angles and moments; and lumbar rotator and bending moments. Our findings may benefit rehabilitation and assessment for return-to-sport 12-24 months post-ACLR. -
G. Athanasopoulos, R. J. Hyndman, N. Kourentzes, and M. O’Hara-Wild, ‘Probabilistic Forecasts Using Expert Judgment: The Road to Recovery From COVID-19,’ Journal of Travel Research, vol. 62, no. 1, pp. 233–258, Jan. 2022, doi: 10.1177/00472875211059240.
Journal Article View at DOI ↗
The COVID-19 pandemic has had a devastating effect on many industries around the world including tourism and policy makers are interested in mapping out what the recovery path will look like. We propose a novel statistical methodology for generating scenario-based probabilistic forecasts based on a large survey of 443 tourism experts and stakeholders. The scenarios map out pessimistic, most-likely and optimistic paths to recovery. Taking advantage of the natural aggregation structure of tourism data due to geographic locations and purposes of travel, we propose combining forecast reconciliation and forecast combinations implemented to historical data to generate robust COVID-free counterfactual forecasts, to contrast against. Our empirical application focuses on Australia, analyzing international arrivals and domestic flows. Both sectors have been severely affected by travel restrictions in the form of international and interstate border closures and regional lockdowns. The two sets of forecasts, allow policy makers to map out the road to recovery and also estimate the expected effect of the pandemic. -
Z. Ghasemi, H. A. Khorshidi, and U. Aickelin, ‘Multi-objective Semi-supervised Clustering for Finding Predictive Clusters’, 2022, arXiv. doi: 10.48550/ARXIV.2201.10764.
Preprint View at DOI ↗
This study concentrates on clustering problems and aims to find compact clusters that are informative regarding the outcome variable. The main goal is partitioning data points so that observations in each cluster are similar and the outcome variable can be predicated using these clusters simultaneously. We model this semi-supervised clustering problem as a multi-objective optimization problem with considering deviation of data points in clusters and prediction error of the outcome variable as two objective functions to be minimized. For finding optimal clustering solutions, we employ a non-dominated sorting genetic algorithm II approach and local regression is applied as prediction method for the output variable. For comparing the performance of the proposed model, we compute seven models using five real-world data sets. Furthermore, we investigate the impact of using local regression for predicting the outcome variable in all models, and examine the performance of the multi-objective models compared to single-objective models. -
M. Abolghasemi, R. J. Hyndman, E. Spiliotis, and C. Bergmeir, ‘Model selection in reconciling hierarchical time series,’ Machine Learning, vol. 111, no. 2, pp. 739–789, Jan. 2022, doi: 10.1007/s10994-021-06126-z.
Preprint View at DOI ↗
-
M. Azizi, U. Aickelin, H. A. Khorshidi, and M. B. Shishehgarkhaneh, ‘Shape and size optimization of truss structures by Chaos game optimization considering frequency constraints,’ Journal of Advanced Research, vol. 41, pp. 89–100, Nov. 2022, doi: 10.1016/j.jare.2022.01.002.
Journal Article View at DOI ↗
INTRODUCTION: An engineering system consists of properly established activities and put together to achieve a predefined goal. These activities include analysis, design, construction, research, and development. Designing and constructing structural systems, including buildings, bridges, highways, and other complex systems, have been developed over the centuries. However, the evolution of these systems has been prolonged because the overall process is very costly and time-consuming, requiring primary human and material resources to be utilized. One of the options for overcoming these shortcomings is the utilization of metaheuristic algorithms as recently developed intelligent techniques. These algorithms can be utilized as upper-level search techniques for optimization procedures to achieve better results. OBJECTIVES: Shape and size optimization of truss structures are considered in this paper utilizing the Chaos Game Optimization (CGO) as one of the recently developed metaheuristic algorithms. The principles of chaos theory and fractal configuration are considered inspirational concepts. METHODS: For the numerical purpose, the 10-bar, 37-bar, 52-bar, 72-bar, and 120-bar truss structures as four of the benchmark problems in this field are considered as design examples in which the frequency constraints are considered as limits that have to be dealt with during the optimization procedure. Multiple optimization runs are also conducted for having a comprehensive statistical analysis, while a comparative investigation is also conducted with other algorithms in the literature. RESULTS: Based on the results of the CGO and other approaches from the literature, the CGO can provide better and competitive results in dealing with the considered truss design problems. CONCLUSION: In summary, the CGO can provide better solutions in dealing with the considered real-size structural design problems with higher levels of complexity. -
A. Ek, A. Schutt, P. J. Stuckey, and G. Tack, ‘Explaining Propagation for Gini and Spread with Variable Mean’, LIPIcs, Volume 235, CP 2022, vol. 235. Schloss Dagstuhl – Leibniz-Zentrum für Informatik, pp. 21:1–21:16, 2022. doi: 10.4230/LIPICS.CP.2022.21.
Preprint View at DOI ↗
In optimisation problems involving multiple agents (stakeholders) we often want to make sure that the solution is balanced and fair. That is, we want to maximise total utility subject to an upper bound on the statistical dispersion (e.g., spread or the Gini coefficient) of the utility given to different agents, or minimise dispersion subject to some lower bounds on utility. These needs arise in, for example, balancing tardiness in scheduling, unwanted shifts in rostering, and desired resources in resource allocation, or minimising deviation from a baseline in schedule repair, to name a few. These problems are often quite challenging. To solve them efficiently we want to effectively reason about dispersion. Previous work has studied the case where the mean is fixed, but this may not be possible for many problems, e.g., scheduling where total utility depends on the final schedule. In this paper we introduce two log-linear-time dispersion propagators - (a) spread (variance, and indirectly standard deviation) and (b) the Gini coefficient - capable of explaining their propagations, thus allowing effective clause learning solvers to be applied to these problems. Propagators for (a) exist in the literature but do not explain themselves, while propagators for (b) have not been previously studied. We avoid introducing floating-point variables, which are usually not supported by learning solvers, by reasoning about scaled, integer versions of the constraints. We show through experimentation that clause learning can substantially improve the solving of problems where we want to bound dispersion and optimise total utility and vice versa. -
Y. Yang, H. Khorshidi, and U. Aickelin, ‘Cluster-based Diversity Over-sampling: A Density and Diversity Oriented Synthetic Over-sampling for Imbalanced Data,’ Proceedings of the 14th International Joint Conference on Computational Intelligence, pp. 17–28, 2022, doi: 10.5220/0011381000003332.
Journal Article View at DOI ↗
-
A. S. Michels and C. G. S. Sikora, ‘A survey on Benders Decomposition methods applied to Assembly Line Balancing Problems,’ IFAC-PapersOnLine, vol. 55, no. 10, pp. 464–469, 2022, doi: 10.1016/j.ifacol.2022.09.437.
Journal Article View at DOI ↗
The assembly line balancing problem (ALBP) assigns tasks to (work)stations to manufacture products. It divides the station loads as evenly as possible since any bottleneck defines the production rate of a flow shop system. The sum of task processing times represents the station load in the classical Simple Assembly Line Balancing Problem (SALBP). Thus, the most loaded station imposes a lower bound on the line's cycle time. However, the simple sum of processing times is only valid under several assumptions. For instance, more realistic ALBPs may contain stochastic data, multiple workers per station, or depend on the production sequence. Hence, the station load computation involves further scheduling decisions for this latter class of problems. In order to tackle these ALBPs with practical extensions, several literature contributions have recently employed Benders decomposition (BD) algorithms. The BD structure allows a division of the formulation into two or more levels. More specifically, this framework permits the decoupling of task assignment decisions and cycle time assessments. In the field of ALBPs, various stochastic and sequence-dependent problems have successfully applied this approach. This paper provides a literature review of these methods, the different formulations, implementation details, and improvement ideas. -
M. Becerra-Fernandez, L. E. Ruiz-Acosta, D. A. Camargo-Mayorga, and M. A. Muñoz, ‘A system dynamics model for sustainable corporate strategic planning’. SciELO journals, 2022. doi: 10.6084/M9.FIGSHARE.20363075.
Dataset View at DOI ↗
Abstract Paper aims This paper presents a manufacturing process model for assessing the effects on economic, social, and environmental targets, given variations on corporate strategies of production, innovation, marketing, and demand for final goods. Originality The model integrates economic, social, and environmental dimensions that are validated through three main scenarios: Business as Usual (no strategic application), Business as Investment (strategic application), and Business as Vision (changes in demand). Research method The model estimates the social, environmental, and economic performance through time based on the System Dynamics methodology. Main findings The results demonstrate the model's suitability as a decision-support tool for sustainability planning in a corporate environment. Implications for theory and practice The model facilitates the analysis of the effects of resource allocation on corporate strategy. -
A. Li, P. Stuckey, S. Koenig, and T. K. S. Kumar, ‘A FastMap-Based Algorithm for Block Modeling,’ Integration of Constraint Programming, Artificial Intelligence, and Operations Research, pp. 232–248, 2022, doi: 10.1007/978-3-031-08011-1_16.
Book Chapter View at DOI ↗
-
H. Bierlee, G. Gange, G. Tack, J. J. Dekker, and P. J. Stuckey, ‘Coupling Different Integer Encodings for SAT,’ Integration of Constraint Programming, Artificial Intelligence, and Operations Research, pp. 44–63, 2022, doi: 10.1007/978-3-031-08011-1_5.
Book Chapter View at DOI ↗
-
P. J. Stuckey and G. Tack, ‘Enumerated Types and Type Extensions for MiniZinc,’ Integration of Constraint Programming, Artificial Intelligence, and Operations Research, pp. 374–389, 2022, doi: 10.1007/978-3-031-08011-1_25.
Book Chapter View at DOI ↗
-
D. Herring, D. Pakravan, and M. Kirley, ‘Analysing Multiobjective Optimization Using Evolutionary Path Length Correlation,’ AI 2021: Advances in Artificial Intelligence, pp. 467–479, 2022, doi: 10.1007/978-3-030-97546-3_38.
Book Chapter View at DOI ↗
2021
-
D. Zhao, H. Wang, J. Huang, and X. Lin, ‘Time-of-Use Pricing for Energy Storage Investment,’ IEEE Transactions on Smart Grid, vol. 13, no. 2, pp. 1165–1177, Mar. 2022, doi: 10.1109/tsg.2021.3136650.
Journal Article View at DOI ↗
Time-of-use (ToU) pricing is widely used by the electricity utility to shave peak load. Such a pricing scheme provides users with incentives to invest in behind-the-meter energy storage and to shift peak load towards low-price intervals. However, without considering the implication on energy storage investment, an improperly designed ToU pricing scheme may lead to significant welfare loss, especially when users over-invest the storage, which leads to new energy consumption peaks. In this paper, we will study how to design a social-optimum ToU pricing scheme by explicitly considering its impact on storage investment. We model the interactions between the utility and users as a two-stage optimization problem. To resolve the challenge of asymmetric information due to users’ private storage cost, we propose a ToU pricing scheme based on different storage types and the aggregate demand per type. Each user does not need to reveal his private cost information. We can further compute the optimal ToU pricing with only a linear complexity. Simulations based on real-world data show that the suboptimality gap of our proposed ToU pricing, compared with the social optimum achieved under complete information, is less than 5%. -
S. Kandanaarachchi, ‘Unsupervised anomaly detection ensembles using item response theory,’ Information Sciences, vol. 587, pp. 142–163, Mar. 2022, doi: 10.1016/j.ins.2021.12.042.
Journal Article View at DOI ↗
-
M. Blom, P. J. Stuckey, V. Teague, and D. Vukcevic, ‘A First Approach to Risk-Limiting Audits for Single Transferable Vote Elections’, arXiv, 2021, doi: 10.48550/ARXIV.2112.09921.
Preprint View at DOI ↗
Risk-limiting audits (RLAs) are an increasingly important method for checking that the reported outcome of an election is, in fact, correct. Indeed, their use is increasingly being legislated. While effective methods for RLAs have been developed for many forms of election -- for example: first-past-the-post, instant-runoff voting, and D'Hondt elections -- auditing methods for single transferable vote (STV) elections have yet to be developed. STV elections are notoriously hard to reason about since there is a complex interaction of votes that change their value throughout the process. In this paper we present the first approach to risk-limiting audits for STV elections, restricted to the case of 2-seat STV elections. -
C. Kermorvant et al., ‘Reconstructing Missing and Anomalous Data Collected from High-Frequency In-Situ Sensors in Fresh Waters,’ International Journal of Environmental Research and Public Health, vol. 18, no. 23, p. 12803, Dec. 2021, doi: 10.3390/ijerph182312803.
Journal Article View at DOI ↗
In situ sensors that collect high-frequency data are used increasingly to monitor aquatic environments. These sensors are prone to technical errors, resulting in unrecorded observations and/or anomalous values that are subsequently removed and create gaps in time series data. We present a framework based on generalized additive and auto-regressive models to recover these missing data. To mimic sporadically missing (i) single observations and (ii) periods of contiguous observations, we randomly removed (i) point data and (ii) day- and week-long sequences of data from a two-year time series of nitrate concentration data collected from Arikaree River, USA, where synoptically collected water temperature, turbidity, conductance, elevation, and dissolved oxygen data were available. In 72% of cases with missing point data, predicted values were within the sensor precision interval of the original value, although predictive ability declined when sequences of missing data occurred. Precision also depended on the availability of other water quality covariates. When covariates were available, even a sudden, event-based peak in nitrate concentration was reconstructed well. By providing a promising method for accurate prediction of missing data, the utility and confidence in summary statistics and statistical trends will increase, thereby assisting the effective monitoring and management of fresh waters and other at-risk ecosystems. -
J. L. Yarmuch, M. Brazil, H. Rubinstein, and D. A. Thomas, ‘A model for open-pit pushback design with operational constraints,’ Optimization and Engineering, Nov. 2021, doi: 10.1007/s11081-021-09699-9.
Journal Article View at DOI ↗
-
A. S. Michels and A. M. Costa, ‘Conserving workforce while temporarily rebalancing assembly lines under demand disruption,’ International Journal of Production Research, vol. 60, no. 21, pp. 6616–6636, Nov. 2021, doi: 10.1080/00207543.2021.1998694.
Journal Article View at DOI ↗
In stable circumstances, assembly lines have workers with different capabilities assigned to stations. They perform a set of specialised tasks multiple times daily. Under a situation of high demand disruption (e.g. the COVID-19 pandemic), the overproduction rate would lead inventory levels to soar. An approach to cope with these demand drops and ongoing workforce costs is to dismiss employees and rebalance the line. Nevertheless, this implies social and economic costs related to rehiring and training. Alternatively, agreements can be made to reduce workload with a proportional wage deduction. These decisions are particularly challenging in heterogeneous workforces. We propose a Mixed-Integer Linear Programming (MILP) model to address the Assembly Line Worker Assignment and Rebalancing Problem (ALWARP). Our model aims at preserving jobs while minimising labour costs. We consider scenarios with falling demands and impose regularity metrics on workload reductions. Computational tests on benchmark datasets show that our strategy can distribute social costs among workers, with only slightly higher cumulative labour hours, while avoiding inconveniences associated with future renovations. A real-world case study of a truck cabin assembly line is investigated: the model can easily incorporate many realistic features and decide which workers should have their workload reduced and at which rate. -
S. Kandanaarachchi and R. J. Hyndman, ‘Leave-One-Out Kernel Density Estimates for Outlier Detection,’ Journal of Computational and Graphical Statistics, vol. 31, no. 2, pp. 586–599, Dec. 2021, doi: 10.1080/10618600.2021.2000425.
Journal Article View at DOI ↗
This article introduces lookout, a new approach to detect outliers using leave-one-out kernel density estimates and extreme value theory. Outlier detection methods that use kernel density estimates generally employ a user defined parameter to determine the bandwidth. Lookout uses persistent homology to construct a bandwidth suitable for outlier detection without any user input. We demonstrate the effectiveness of lookout on an extensive data repository by comparing its performance with other outlier detection methods based on extreme value theory. Furthermore, we introduce outlier persistence, a useful concept that explores the birth and the cessation of outliers with changing bandwidth and significance levels. The R package lookout implements this algorithm. Supplementary files for this article are available online. -
N. Andrés-Thió, M. Brazil, C. Ras, D. Thomas, and M. Volz, ‘An exact algorithm for constructing minimum Euclidean skeletons of polygons,’ Journal of Global Optimization, vol. 83, no. 1, pp. 137–162, Oct. 2021, doi: 10.1007/s10898-021-01101-3.
Journal Article View at DOI ↗
-
M. Volz, M. Brazil, C. Ras, and D. Thomas, ‘Simplifying obstacles for Steiner network problems in the plane,’ Networks, vol. 80, no. 1, pp. 77–92, Oct. 2021, doi: 10.1002/net.22080.
Journal Article View at DOI ↗
Abstract We present methods for simplifying the geometry of polygonal obstacles as a preprocessing step to solving obstacle‐avoiding Steiner network problems in the plane. The methods reduce the total number of vertices and edges that need to be considered for the given obstacles, and their use is expected to significantly improve the efficiency of exact algorithms for solving a range of practical Steiner network problems in obstacle environments. Included are methods for extending obstacles (via a new padding method and a backfilling procedure from the literature), and various methods for simplifying obstacles, including new methods called bounding and eliminating . We show that these methods reduce the total number of obstacle vertices and edges by performing experiments on obstacles with up to 100 vertices in the presence of up to 100 terminals. The experiments utilize a modified version of a known algorithm for quickly generating large numbers of “random” polygons with hundreds of vertices. Corresponding datasets and implementations have been made available on GitHub. -
K. Leo, C. Mears, G. Tack, and M. Garcia de la Banda, ‘Globalizing constraint models,’ Artificial Intelligence, vol. 302, p. 103599, Jan. 2022, doi: 10.1016/j.artint.2021.103599.
Journal Article View at DOI ↗
-
T. C. Lopes, A. S. Michels, C. G. S. Sikora, N. Brauner, and L. Magatão, ‘Assembly line balancing for two cycle times: Anticipating demand fluctuations,’ Computers & Industrial Engineering, vol. 162, p. 107685, Dec. 2021, doi: 10.1016/j.cie.2021.107685.
Journal Article View at DOI ↗
-
D. Whittle, M. Brazil, P. A. Grossman, J. H. Rubinstein, and D. A. Thomas, ‘Minimum Steiner trees on a set of concyclic points and their center,’ International Transactions in Operational Research, vol. 29, no. 4, pp. 2201–2225, Sep. 2021, doi: 10.1111/itor.13055.
Journal Article View at DOI ↗
Abstract Consider a configuration of points comprising a point q and a set of concyclic points R that are all a given distance r from q in the Euclidean plane. In this paper, we investigate the relationship between the length of a minimum Steiner tree (MStT) on and a minimum spanning tree on R . We show that if the degree of q in the MStT is 1, then the difference between these two lengths is at least , and that this lower bound is tight. This bound can be applied as part of an efficient algorithm to find the solution to the prize‐collecting Euclidean Steiner tree problem, as outlined in an earlier paper. -
A. De Coster, N. Musliu, A. Schaerf, J. Schoisswohl, and K. Smith-Miles, ‘Algorithm selection and instance space analysis for curriculum-based course timetabling,’ Journal of Scheduling, vol. 25, no. 1, pp. 35–58, Sep. 2021, doi: 10.1007/s10951-021-00701-x.
Journal Article View at DOI ↗
-
F. J. Aragón Artacho, R. Campoy, and M. K. Tam, ‘Strengthened splitting methods for computing resolvents,’ Computational Optimization and Applications, vol. 80, no. 2, pp. 549–585, Aug. 2021, doi: 10.1007/s10589-021-00291-6.
Journal Article View at DOI ↗
-
E. R. Csetnek, A. Eberhard, and M. K. Tam, ‘Convergence rates for boundedly regular systems,’ Advances in Computational Mathematics, vol. 47, no. 5, Aug. 2021, doi: 10.1007/s10444-021-09891-6.
Preprint View at DOI ↗
-
H. Hewamalage, P. Montero-Manso, C. Bergmeir, and R. J. Hyndman, ‘A Look at the Evaluation Setup of the M5 Forecasting Competition’, 2021, arXiv. doi: 10.48550/ARXIV.2108.03588.
Preprint View at DOI ↗
Forecast evaluation plays a key role in how empirical evidence shapes the development of the discipline. Domain experts are interested in error measures relevant for their decision making needs. Such measures may produce unreliable results. Although reliability properties of several metrics have already been discussed, it has hardly been quantified in an objective way. We propose a measure named Rank Stability, which evaluates how much the rankings of an experiment differ in between similar datasets, when the models and errors are constant. We use this to study the evaluation setup of the M5. We find that the evaluation setup of the M5 is less reliable than other measures. The main drivers of instability are hierarchical aggregation and scaling. Price-weighting reduces the stability of all tested error measures. Scale normalization of the M5 error measure results in less stability than other scale-free errors. Hierarchical levels taken separately are less stable with more aggregation, and their combination is even less stable than individual levels. We also show positive tradeoffs of retaining aggregation importance without affecting stability. Aggregation and stability can be linked to the influence of much debated magic numbers. Many of our findings can be applied to general hierarchical forecast benchmarking. -
E. Spiliotis, M. Abolghasemi, R. J. Hyndman, F. Petropoulos, and V. Assimakopoulos, ‘Hierarchical forecast reconciliation with machine learning,’ Applied Soft Computing, vol. 112, p. 107756, Nov. 2021, doi: 10.1016/j.asoc.2021.107756.
Preprint View at DOI ↗
-
K. Bandara, R. J. Hyndman, and C. Bergmeir, ‘MSTL: A Seasonal-Trend Decomposition Algorithm for Time Series with Multiple Seasonal Patterns’, 2021, arXiv. doi: 10.48550/ARXIV.2107.13462.
Preprint View at DOI ↗
The decomposition of time series into components is an important task that helps to understand time series and can enable better forecasting. Nowadays, with high sampling rates leading to high-frequency data (such as daily, hourly, or minutely data), many real-world datasets contain time series data that can exhibit multiple seasonal patterns. Although several methods have been proposed to decompose time series better under these circumstances, they are often computationally inefficient or inaccurate. In this study, we propose Multiple Seasonal-Trend decomposition using Loess (MSTL), an extension to the traditional Seasonal-Trend decomposition using Loess (STL) procedure, allowing the decomposition of time series with multiple seasonal patterns. In our evaluation on synthetic and a perturbed real-world time series dataset, compared to other decomposition benchmarks, MSTL demonstrates competitive results with lower computational cost. The implementation of MSTL is available in the R package forecast. -
A. Zamani, H. Haghbin, M. Hashemi, and R. J. Hyndman, ‘Seasonal functional autoregressive models,’ Journal of Time Series Analysis, vol. 43, no. 2, pp. 197–218, Aug. 2021, doi: 10.1111/jtsa.12608.
Preprint View at DOI ↗
Functional autoregressive models are popular for functional time series analysis, but the standard formulation fails to address seasonal behaviour in functional time series data. To overcome this shortcoming, we introduce seasonal functional autoregressive time series models. For the model of order one, we derive sufficient stationarity conditions and limiting behaviour, and provide estimation and prediction methods. Moreover, we consider a portmanteau test for testing the adequacy of this model, and we derive its asymptotic distribution. The merits of this model are demonstrated using simulation studies and via an application to hourly pedestrian counts. -
E. Yap, M. A. Muñoz, and K. Smith-Miles, ‘On the diversity and robustness of parameterised multi-objective test suites,’ Applied Soft Computing, vol. 110, p. 107613, Oct. 2021, doi: 10.1016/j.asoc.2021.107613.
Journal Article View at DOI ↗
-
M. Ashouri, R. J. Hyndman, and G. Shmueli, ‘Fast Forecast Reconciliation Using Linear Models,’ Journal of Computational and Graphical Statistics, vol. 31, no. 1, pp. 263–282, Jul. 2021, doi: 10.1080/10618600.2021.1939038.
Preprint View at DOI ↗
Forecasting hierarchical or grouped time series using a reconciliation approach involves two steps: computing base forecasts and reconciling the forecasts. Base forecasts can be computed by popular time series forecasting methods such as exponential smoothing (ETS) and Autoregressive Integrated Moving Average (ARIMA) models. The reconciliation step is a linear process that adjusts the base forecasts to ensure they are coherent. However, using ETS or ARIMA for base forecasts can be computationally challenging when there are a large number of series to forecast, as each model must be numerically optimized for each series. We propose a linear model that avoids this computational problem and handles the forecasting and reconciliation in a single step. The proposed method is very flexible in incorporating external data. We illustrate our approach using a dataset on monthly Australian domestic tourism, as well as a simulated dataset. We compare our approach to reconciliation using ETS and ARIMA, and show that our approach is much faster while providing similar levels of forecast accuracy. Supplementary files for this article are available online. -
S. Gupta, R. J. Hyndman, D. Cook, and A. Unwin, ‘Visualizing Probability Distributions Across Bivariate Cyclic Temporal Granularities,’ Journal of Computational and Graphical Statistics, vol. 31, no. 1, pp. 14–25, Jul. 2021, doi: 10.1080/10618600.2021.1938588.
Journal Article View at DOI ↗
Deconstructing a time index into time granularities can assist in exploration and automated analysis of large temporal datasets. This article describes classes of time deconstructions using linear and cyclic time granularities. Linear granularities respect the linear progression of time such as hours, days, weeks and months. Cyclic granularities can be circular such as hour-of-the-day, quasi-circular such as day-of-the-month, and aperiodic such as public holidays. The hierarchical structure of granularities creates a nested ordering: hour-of-the-day and second-of-the-minute are single-order-up. Hour-of-the-week is multiple-order-up, because it passes over day-of-the-week. Methods are provided for creating all possible granularities for a time index. A recommendation algorithm provides an indication whether a pair of granularities can be meaningfully examined together (a “harmony”), or when they cannot (a “clash”). Time granularities can be used to create data visualizations to explore for periodicities, associations and anomalies. The granularities form categorical variables (ordered or unordered) which induce groupings of the observations. Assuming a numeric response variable, the resulting graphics are then displays of distributions compared across combinations of categorical variables. The methods implemented in the open source R package gravitas are consistent with a tidy workflow, with probability distributions examined using the range of graphics available in ggplot2. Supplementary files for this article are available online. -
M. N. Dao, N. D. Dizon, J. A. Hogan, and M. K. Tam, ‘Constraint Reduction Reformulations for Projection Algorithms with Applications to Wavelet Construction,’ Journal of Optimization Theory and Applications, vol. 190, no. 1, pp. 201–233, Jun. 2021, doi: 10.1007/s10957-021-01878-z.
Journal Article View at DOI ↗
-
J. Devriendt, S. Gocht, E. Demirović, J. Nordström, and P. J. Stuckey, ‘Cutting to the Core of Pseudo-Boolean Optimization: Combining Core-Guided Search with Cutting Planes Reasoning,’ Proceedings of the AAAI Conference on Artificial Intelligence, vol. 35, no. 5, pp. 3750–3758, May 2021, doi: 10.1609/aaai.v35i5.16492.
Journal Article View at DOI ↗
Core-guided techniques have revolutionized Boolean satisfiability approaches to optimization problems (MaxSAT), but the process at the heart of these methods, strengthening bounds on solutions by repeatedly adding cardinality constraints, remains a bottleneck. Cardinality constraints require significant work to be re-encoded to SAT, and SAT solvers are notoriously weak at cardinality reasoning. In this work, we lift core-guided search to pseudo-Boolean (PB) solvers, which deal with more general PB optimization problems and operate natively with cardinality constraints. The cutting planes method used in such solvers allows us to derive stronger cardinality constraints, which yield better updates to solution bounds, and the increased efficiency of objective function reformulation also makes it feasible to switch repeatedly between lower-bounding and upper- bounding search. A thorough evaluation on applied and crafted benchmarks shows that our core-guided PB solver significantly improves on the state of the art in pseudo-Boolean optimization. -
C. Cheng, R. Zhu, A. M. Costa, R. G. Thompson, and X. Huang, ‘Multi-period two-echelon location routing problem for disaster waste clean-up,’ Transportmetrica A: Transport Science, vol. 18, no. 3, pp. 1053–1083, May 2021, doi: 10.1080/23249935.2021.1916644.
Journal Article View at DOI ↗
Waste clean-up after a disaster is one of the most critical tasks in the response stage of disaster management. We develop a model to minimise the cost and duration of disaster waste clean-up considering using Temporary Disaster Waste Management Sites (TDWMSs), which can store and process waste before it is sent to the final disposal sites. The problem that arises can be seen as a Multi-Period Two-echelon Location Routing Problem (MP-2ELRP) in which the main decisions are the location of the TDWMSs and the routing of vehicles in both echelons. In this paper, we propose both a mixed-integer program and a Genetic Algorithm (GA) to model and solve the problem. Computational tests indicate: (i) the performance of proposed GA is robust; (ii) the use of TDWMSs can reduce both total waste clean-up cost and duration; and (iii) the capacities of TDWMSs have a significant impact on the total waste clean-up time and duration. -
B. Rostami-Tabar, M. M. Ali, T. Hong, R. J. Hyndman, M. D. Porter, and A. Syntetos, ‘Forecasting for social good,’ International Journal of Forecasting, vol. 38, no. 3, pp. 1245–1257, Jul. 2022, doi: 10.1016/j.ijforecast.2021.02.010.
Preprint View at DOI ↗
-
C. M. Baker, I. Chades, J. McVernon, A. Robinson, and H. Bondell, ‘Optimal allocation of PCR tests to minimise disease transmission through contact tracing and quarantine,’ Mar. 2021, doi: 10.1101/2021.03.23.21254148.
Preprint View at DOI ↗
Abstract PCR testing is a crucial capability for managing disease outbreaks, but it is also a limited resource and must be used carefully to ensure the information gain from testing is valuable. Testing has two broad uses, namely to track epidemic dynamics and to reduce transmission by identifying and managing cases. In this work we develop a modelling framework to examine the effects of test allocation in an epidemic, with a focus on using testing to minimise transmission. Using the COVID-19 pandemic as an example, we examine how the number of tests conducted per day relates to reduction in disease transmission, in the context of logistical constraints on the testing system. We show that if daily testing is above the routine capacity of a testing system, which can cause delays, then those delays can undermine efforts to reduce transmission through contact tracing and quarantine. This work highlights that the two goals of aiming to reduce transmission and aiming to identify all cases are different, and it is possible that focusing on one may undermine achieving the other. To develop an effective strategy, the goals must be clear and performance metrics must match the goals of the testing strategy. If metrics do not match the objectives of the strategy, then those metrics may incentivise actions that undermine achieving the objectives. -
L. H. d. S. Fernandes, A. C. Lorena, and K. Smith-Miles, ‘Towards Understanding Clustering Problems and Algorithms: An Instance Space Analysis,’ Algorithms, vol. 14, no. 3, p. 95, Mar. 2021, doi: 10.3390/a14030095.
Journal Article View at DOI ↗
Various criteria and algorithms can be used for clustering, leading to very distinct outcomes and potential biases towards datasets with certain structures. More generally, the selection of the most effective algorithm to be applied for a given dataset, based on its characteristics, is a problem that has been largely studied in the field of meta-learning. Recent advances in the form of a new methodology known as Instance Space Analysis provide an opportunity to extend such meta-analyses to gain greater visual insights of the relationship between datasets’ characteristics and the performance of different algorithms. The aim of this study is to perform an Instance Space Analysis for the first time for clustering problems and algorithms. As a result, we are able to analyze the impact of the choice of the test instances employed, and the strengths and weaknesses of some popular clustering algorithms, for datasets with different structures. -
P. B. Castellucci, A. M. Costa, and F. Toledo, ‘Network scheduling problem with cross-docking and loading constraints,’ Computers & Operations Research, vol. 132, p. 105271, Aug. 2021, doi: 10.1016/j.cor.2021.105271.
Journal Article View at DOI ↗
-
J. L. Yarmuch, M. Brazil, H. Rubinstein, and D. A. Thomas, ‘A mathematical model for mineable pushback designs,’ International Journal of Mining, Reclamation and Environment, vol. 35, no. 7, pp. 523–539, Feb. 2021, doi: 10.1080/17480930.2021.1885582.
Journal Article View at DOI ↗
In the design of open pit mines, the region to be mined is partitioned into pushbacks, subregions that allow the mining to be divided into distinct phases. Practical pushbacks are connected, satisfy a minimum width for mining equipment and include a haulage ramp. Current pushback models typically relax some or all of these mineability conditions; consequently, the outputs from those models require significant intervention by mining engineers. We present a formulation to generate maximum value practical pushbacks. A closeness factor is introduced to quantify the design’s mineability. Finally, a case study of a real mine shows that our model can produce pushbacks with more practical designs and better value than traditional approaches. -
M. A. Muñoz, M. Kirley, and K. Smith-Miles, ‘Analyzing randomness effects on the reliability of exploratory landscape analysis,’ Natural Computing, vol. 21, no. 2, pp. 131–154, Feb. 2021, doi: 10.1007/s11047-021-09847-1.
Journal Article View at DOI ↗
-
I. Senthooran, P. Le Bodic, and P. J. Stuckey, ‘Optimising Training for Service Delivery’, LIPIcs, Volume 210, CP 2021, vol. 210. Schloss Dagstuhl – Leibniz-Zentrum für Informatik, pp. 48:1–48:15, 2021. doi: 10.4230/LIPICS.CP.2021.48.
Journal Article View at DOI ↗
We study the problem of training a roster of engineers, who are scheduled to respond to service calls that require a set of skills, and where engineers and calls have different locations. Both training an engineer in a skill and sending an engineer to respond a non-local service call incur a cost. Alternatively, a local contractor can be hired. The problem consists in training engineers in skills so that the quality of service (i.e. response time) is maximised and costs are minimised. The problem is hard to solve in practice partly because (1) the value of training an engineer in one skill depends on other training decisions, (2) evaluating training decisions means evaluating the schedules that are now made possible by the new skills, and (3) these schedules must be computed over a long time horizon, otherwise training may not pay off. We show that a monolithic approach to this problem is not practical. Instead, we decompose it into three subproblems, modelled with MiniZinc. This allows us to pick the approach that works best for each subproblem (MIP or CP) and provide good solutions to the problem. Data is provided by a multinational company. -
H. Alipour, M. A. Munoz Acosta, and K. Smith-Miles, ‘Instance Space Analysis for the Maximum Flow Problem’. figshare, 2021. doi: 10.6084/M9.FIGSHARE.14761836.V1.
Dataset View at DOI ↗
Metadata and source codes to explore the instance space of the maximum flow problem online.
2020
-
M. K. Tam, ‘Gearhart–Koshy acceleration for affine subspaces,’ Operations Research Letters, vol. 49, no. 2, pp. 157–163, Mar. 2021, doi: 10.1016/j.orl.2020.12.007.
Preprint View at DOI ↗
-
K. Smith-Miles, J. Christiansen, and M. A. Muñoz, ‘Revisiting where are the hard knapsack problems? via Instance Space Analysis,’ Computers & Operations Research, vol. 128, p. 105184, Apr. 2021, doi: 10.1016/j.cor.2020.105184.
Journal Article View at DOI ↗
-
K. Smith-Miles and X. Geng, ‘Revisiting Facial Age Estimation With New Insights From Instance Space Analysis,’ IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 5, pp. 2689–2697, May 2022, doi: 10.1109/tpami.2020.3038760.
Journal Article View at DOI ↗
When demonstrating the effectiveness of a new algorithm, researchers are traditionally encouraged to compare their algorithm's performance against existing algorithms on well-studied benchmark test suites. In the absence of more nuanced methodologies, algorithm performance is typically summarized on average across the test suite examples. This paper highlights the potential bias of conclusions drawn by analyzing "on average" performance, and the opportunities offered by a recent testing methodology known as instance space analysis. To illustrate, we revisit our 2007 comparative study of algorithms for facial age estimation, and rigorously stress-test to challenge the original conclusions. The case study demonstrates how powerful visualizations offered by instance space analysis enable greater insights into unique strengths and weaknesses, and which algorithm should be used when and why. Inspired by such insights, a new algorithm is proposed, and its unique advantage is demonstrated. The bias often hidden in well-studied datasets, and the ramifications for drawing biased conclusions, are also illustrated in this case study. While focused on facial age estimation, the methodology and lessons learned from the case study are broadly applicable to any study seeking to draw conclusions about algorithm performance based on empirical results. -
C. J. Ras, M. Brazil, and D. A. Thomas, ‘Computational complexity of the 2-connected Steiner network problem in the ℓ plane,’ Theoretical Computer Science, vol. 850, pp. 168–184, Jan. 2021, doi: 10.1016/j.tcs.2020.11.002.
Journal Article View at DOI ↗
-
C. Cheng, R. Zhu, A. M. Costa, and R. G. Thompson, ‘Optimisation of waste clean-up after large-scale disasters,’ Waste Management, vol. 119, pp. 1–10, Jan. 2021, doi: 10.1016/j.wasman.2020.09.023.
Journal Article View at DOI ↗
-
A. S. Michels and A. M. Costa, ‘A note to: A multiple-rule based constructive randomized search algorithm for solving assembly line worker assignment and balancing problem,’ Journal of Intelligent Manufacturing, vol. 32, no. 8, pp. 2121–2124, Jul. 2020, doi: 10.1007/s10845-020-01632-8.
Journal Article View at DOI ↗
-
Y. Kang, R. J. Hyndman, and F. Li, ‘GRATIS: GeneRAting TIme Series with diverse and controllable characteristics,’ Statistical Analysis and Data Mining: The ASA Data Science Journal, vol. 13, no. 4, pp. 354–376, May 2020, doi: 10.1002/sam.11461.
Journal Article View at DOI ↗
Abstract The explosion of time series data in recent years has brought a flourish of new time series analysis methods, for forecasting, clustering, classification and other tasks. The evaluation of these new methods requires either collecting or simulating a diverse set of time series benchmarking data to enable reliable comparisons against alternative approaches. We propose GeneRAting TIme Series with diverse and controllable characteristics, named GRATIS, with the use of mixture autoregressive (MAR) models. We simulate sets of time series using MAR models and investigate the diversity and coverage of the generated time series in a time series feature space. By tuning the parameters of the MAR models, GRATIS is also able to efficiently generate new time series with controllable features. In general, as a costless surrogate to the traditional data collection approach, GRATIS can be used as an evaluation tool for tasks such as time series forecasting and classification. We illustrate the usefulness of our time series generation process through a time series forecasting application. -
A. Ek, M. Garcia de la Banda, A. Schutt, P. J. Stuckey, and G. Tack, ‘Modelling and Solving Online Optimisation Problems,’ Proceedings of the AAAI Conference on Artificial Intelligence, vol. 34, no. 02, pp. 1477–1485, Apr. 2020, doi: 10.1609/aaai.v34i02.5506.
Journal Article View at DOI ↗
Many optimisation problems are of an online—also called dynamic—nature, where new information is expected to arrive and the problem must be resolved in an ongoing fashion to (a) improve or revise previous decisions and (b) take new ones. Typically, building an online decision-making system requires substantial ad-hoc coding to ensure the offline version of the optimisation problem is continually adjusted and resolved. This paper defines a general framework for automatically solving online optimisation problems. This is achieved by extending a model of the offline optimisation problem, from which an online version is automatically constructed, thus requiring no further modelling effort. In doing so, it formalises many of the aspects that arise in online optimisation problems. The same framework can be applied for automatically creating sliding-window solving approaches for problems that have a large time horizon. Experiments show we can automatically create efficient online and sliding-window solutions to optimisation problems.