Jan Harm van Vuuren
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jan Harm van Vuuren has a total of 55 co-authors. Here are the top co-authors by number of shared publications:
- Alewyn P. Burger
University of Stellenbosch
4 shared publications - Darian Raad
Ph.D. Candidate, Dept. of Logistics, Univ. of Stellenbosch, Private Bag X1, Matieland 7602, South Africa
3 shared publications - Louis Wilhelm Ernst Boshoff
Stellenbosch University
3 shared publications AP Burger
2 shared publications- Alewyn P. Burger
Department of Mathematics and Statistics, University of Victoria, P.O. Box 3045, Victoria, BC, Canada, V8W 3P4
2 shared publications - Alexander Sinske
GLS Consulting, Stellenbosch 7600, South Africa
2 shared publications
Latest Publications
- Open Access
- 9 September 2026
- Published by Taylor & Francis in International Journal of Production Research
- p. 1-48
A new school bus location routing problem with mixed loads and bus stop selection
1. The school bus routing problem (SBRP) is a complex, highly relevant class of vehicle routing problems (VRPs) that is concerned with finding an optimal pickup and drop-off sequence for students f...
- 9 September 2026
- Published by Taylor & Francis in International Journal of Production Research
- p. 1-48
Open AccessA new school bus location routing problem with mixed loads and bus stop selection
1. The school bus routing problem (SBRP) is a complex, highly relevant class of vehicle routing problems (VRPs) that is concerned with finding an optimal pickup and drop-off sequence for students f...
- Open Access
- 6 September 2026
- Published by Elsevier in Next Research
- Vol. 13
Network-informed sentiment estimation under sparse financial-news coverage
- 6 September 2026
- Published by Elsevier in Next Research
- Vol. 13
Open AccessNetwork-informed sentiment estimation under sparse financial-news coverage
- Open Access
- 30 July 2026
- Published by Elsevier in SSRN Electronic Journal
A fitness landscape-integrated, problem-class agnostic metaheuristic selection & configuration framework for binary programming problems
There has recently been a remarkable increase in the number of metaheuristic optimisation methods proposed in the literature. As a result, researchers and analysts face the challenging and time-consuming task of identifying an appropriate metaheuristic from the vast array of approximate solution methodologies available. In this paper, we propose a metaheuristic selection and configuration framework for reducing the effects of performance complementarity by integrating solutions to the problems of algorithm selection and algorithm configuration in a simple and efficient manner within the application context of binary programming problems. This framework operates in two phases --- an offline phase and an online phase. The offline phase takes as input a set of binary programming problem instances and a user-specified algorithm portfolio, and guides the user through constructing a comprehensive and coherent meta-learning database {and training a suite of tailored performance prediction models}. Building on the {computational foundation} of the offline phase, the online phase is tailored to take as input unseen binary programming problem instances and {rapidly} produce {bi-objective} metaheuristic configuration recommendations {that approximate the true Pareto front} of algorithmic configurations that achieve trade-offs between minimising computation time and maximising solution quality. The framework is applied in a problem class-agnostic manner to benchmark instances in four well-known binary programming problem classes: The knapsack problem, the travelling salesperson problem, the graph vertex colouring problem, and the one-dimensional bin packing problem. The framework is finally validated by applying it to a real-world-inspired context involving a novel plant breeding binary programming problem.
- 30 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessA fitness landscape-integrated, problem-class agnostic metaheuristic selection & configuration framework for binary programming problems
There has recently been a remarkable increase in the number of metaheuristic optimisation methods proposed in the literature. As a result, researchers and analysts face the challenging and time-consuming task of identifying an appropriate metaheuristic from the vast array of approximate solution methodologies available. In this paper, we propose a metaheuristic selection and configuration framework for reducing the effects of performance complementarity by integrating solutions to the problems of algorithm selection and algorithm configuration in a simple and efficient manner within the application context of binary programming problems. This framework operates in two phases --- an offline phase and an online phase. The offline phase takes as input a set of binary programming problem instances and a user-specified algorithm portfolio, and guides the user through constructing a comprehensive and coherent meta-learning database {and training a suite of tailored performance prediction models}. Building on the {computational foundation} of the offline phase, the online phase is tailored to take as input unseen binary programming problem instances and {rapidly} produce {bi-objective} metaheuristic configuration recommendations {that approximate the true Pareto front} of algorithmic configurations that achieve trade-offs between minimising computation time and maximising solution quality. The framework is applied in a problem class-agnostic manner to benchmark instances in four well-known binary programming problem classes: The knapsack problem, the travelling salesperson problem, the graph vertex colouring problem, and the one-dimensional bin packing problem. The framework is finally validated by applying it to a real-world-inspired context involving a novel plant breeding binary programming problem.
- Open Access
- 5 May 2026
- Published by Elsevier in Socio-Economic Planning Sciences
- Vol. 105
Integrating food insecurity mapping and equitable resource allocation for foodbank redistribution in an end-to-end decision support framework
- 5 May 2026
- Published by Elsevier in Socio-Economic Planning Sciences
- Vol. 105
Open AccessIntegrating food insecurity mapping and equitable resource allocation for foodbank redistribution in an end-to-end decision support framework
- Open Access
- 2 February 2026
- Published by Elsevier in SSRN Electronic Journal
Beyond peripheries: A conditional, sector-based approach to portfolio selection
Financial market networks (FMNs) offer a promising alternative to mean-variance (MV) optimisation, yet current methodologies typically bifurcate into unconditional peripheral selection for diversification or conditional selection based on centrality-performance trade-offs. This study bridges these paradigms by proposing sector-conditional optimal selection (SCOS), a hybrid framework that enforces structural diversification through economic partitioning while applying a local, correlation-based selection rule within each sector. By analysing the alignment between centrality and asset performance, the SCOS mechanism dynamically switches between peripheral assets and structural hubs. Empirical validation on the NSE 500 (2012-2015) demonstrates that the SCOS strategy achieves a cost-adjusted Sharpe ratio of 1.47, outperforming both the naive market benchmark (1.41) and the notorious MV-strategy (1.29), while significantly mitigating maximum drawdown. Monte-Carlo simulations confirm the statistical robustness of the SCOS strategy, and a Markov regime analysis reveals that the approach generates alpha by preserving defensive peripheral exposure during dominant low-coupling regimes while opportunistically exploiting transient high-coupling structural shifts.
- 2 February 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessBeyond peripheries: A conditional, sector-based approach to portfolio selection
Financial market networks (FMNs) offer a promising alternative to mean-variance (MV) optimisation, yet current methodologies typically bifurcate into unconditional peripheral selection for diversification or conditional selection based on centrality-performance trade-offs. This study bridges these paradigms by proposing sector-conditional optimal selection (SCOS), a hybrid framework that enforces structural diversification through economic partitioning while applying a local, correlation-based selection rule within each sector. By analysing the alignment between centrality and asset performance, the SCOS mechanism dynamically switches between peripheral assets and structural hubs. Empirical validation on the NSE 500 (2012-2015) demonstrates that the SCOS strategy achieves a cost-adjusted Sharpe ratio of 1.47, outperforming both the naive market benchmark (1.41) and the notorious MV-strategy (1.29), while significantly mitigating maximum drawdown. Monte-Carlo simulations confirm the statistical robustness of the SCOS strategy, and a Markov regime analysis reveals that the approach generates alpha by preserving defensive peripheral exposure during dominant low-coupling regimes while opportunistically exploiting transient high-coupling structural shifts.
- 28 October 2025
- Published by Springer Nature
- p. 16-30
Aligning Metrics in Time Series Anomaly Detection with Desirable Behaviour in Practice
- 28 October 2025
- Published by Springer Nature
- p. 16-30
Aligning Metrics in Time Series Anomaly Detection with Desirable Behaviour in Practice
- 28 October 2025
- Published by Springer Nature
- p. 224-237
A Knowledge Graph Approach Towards Detecting Large Language Model Hallucination
- 28 October 2025
- Published by Springer Nature
- p. 224-237
A Knowledge Graph Approach Towards Detecting Large Language Model Hallucination
- 28 October 2025
- Published by Springer Nature
- p. 105-118
Transformer-Based Decision Support for Multi-Objective Recommendations
- 28 October 2025
- Published by Springer Nature
- p. 105-118
Transformer-Based Decision Support for Multi-Objective Recommendations
- 28 October 2025
- Published by Springer Nature
- p. 43-56
An Efficient BERT-Based Algorithm for Extracting Opinion Targets from Financial News
- 28 October 2025
- Published by Springer Nature
- p. 43-56
An Efficient BERT-Based Algorithm for Extracting Opinion Targets from Financial News
- 11 October 2025
- Published by Springer Nature
- p. 37-47
The Cost-Benefit Trade-Off Between Problem-Specific and Algorithm-Specific Meta-learning for the Knapsack Problem
- 11 October 2025
- Published by Springer Nature
- p. 37-47
The Cost-Benefit Trade-Off Between Problem-Specific and Algorithm-Specific Meta-learning for the Knapsack Problem