Andries Heyns | Hanken School of Economics (original) (raw)

Papers by Andries Heyns

Transportation

The traditional aim in transportation planning is to maximise gains associated with vehicular tra... more The traditional aim in transportation planning is to maximise gains associated with vehicular travel distances or times, indirectly prioritising populations that live near existing or proposed roads - remote populations that first require hours of walking to reach roads are overlooked. In this paper, rural roads optimisation is performed using a new model that estimates proposed roads' accessibility gains, considering reductions in vehicular travel time and reductions in walking time required by remote populations to reach them. This ensures that even the most remote populations that may benefit from new roads are included in their evaluation. When presented with a large number of proposed roads and the requirement of determining a plan within a suitable budget, it is often infeasible to construct all proposed roads. In such instances, subsets of well-performing road-combinations that are evaluated with respect to multiple objectives need to be identified for analysis and comparison-for which multi-objective optimisation approaches can be employed. Traditional optimisation approaches return a small number of road-combination plans only, limited to user-specified budget levels and objective weight sets. This paper presents an innovative heuristic solution approach that overcomes such limitations by returning thousands of well-performing solutions scattered across a budget span, and not limited in number to user-specified objective weight sets at fixed budget levels. The heuristic is employed along with a more traditional weighted-sum integer-linear programming approach to determine high-quality road-combination plans selected from 92 roads recently proposed for construction in Nepal's remote Karnali province. Using these two approaches with inputs from the new multi-modal accessibility model, it is illustrated how rural roads planning can be performed to the benefit of rural populations regardless of their proximity to roads. New planning and analysis benefits of the heuristic are demonstrated by comparing its solutions to those determined by the weighted-sum approach, providing a level of detail and sophistication not previously possible for rural roads planning and analysis.

After 23 years [1997-2019] deploying wildfire detection cameras for both forestry corporates and ... more After 23 years [1997-2019] deploying wildfire detection cameras for both forestry corporates and government agencies this study aims to assess improvements related to IFM and the associated trends in reduced damage over the last 18 years [2001-2018]. Satellite based burn scar measurements in the area under camera detection have picked up reduced wildfire damage overall for 18 years. Each calendar month also shows an average reduction over the period except for July where fire break burns have been increasing, and November which is evidence if the increased duration of the winter fire season. All other provinces in the country have seen increased burn frequencies measured using the same satellite data over the same period. Optimizing wildfire detection using numerical methods to maximize line-of-sight coverage for cameras placed at high points is reported here. Introducing real word constraints like site access and connectivity requirements for linking all the detection high sites (c...

Expert Systems With Applications, 2020

Frontiers in Ecology and Evolution, 2020

Journal of Transport Geography, 2021

Transport and economic geographers' accessibility models provide nuanced descriptions of accessib... more Transport and economic geographers' accessibility models provide nuanced descriptions of accessibility gains to investments in structured environments but struggle to accurately reflect the complex mobility constraints in rural, mountainous areas of the developing world. This forces planners in such contexts to employ cruder measures of accessibility gains to each road, particularly where large collections of feeder roads inflate data collection expenses. To address this disconnect we develop a scalable method for evaluating the cost-efficiency of rural roads investments based on seasonal accessibility improvements to specified services. Accessibility improvements are measured using road-specific multi-modal cost-distance models, incorporating terrain, seasonal effects, and extensive off-network walking travel. Using our models, we estimate and compare accessibility improvements for a large collection of proposed rural feeder roads in a case study of Nepal's remote, mountainous Karnali province. The developed model and workflow can be adapted to compare accessibility gains from roads or other investments in equivalently rugged, remote, and data-poor environments, opening the door to more rigorous accounting of accessibility in a traditionally neglected context.

Computers, Environment & Urban Systems, 2021

The optimisation of surveillance and detection systems comprised of specialised cameras is a well... more The optimisation of surveillance and detection systems comprised of specialised cameras is a well-known problem in the operations research literature. In these problems, the aim is to locate optimal camera sites so that their combined coverage with respect to some area of interest-called a cover zone-is maximised. The standard approach is to maximise cover with respect to a single cover zone, and to consider either cameras providing rotational (360 •) cover, or cameras fixed to a specific direction and with visibility limited to within the camera's field-of-view. The Rhino Pride Foundation in South Africa required the optimisation of a camera surveillance system for a new protected area. Their coverage requirements were, however, beyond what has been previously encountered in the literature. Four covering objectives over three separate cover zones were to be maximised, while the system was to be optimised for rotational cover during the day, and some cameras would be required to be fixed towards a high-risk zone at night and limited to their field-of-view. A novel multi-attribute genetic algorithm based on the popular NSGA-II was developed for this purpose. Various solutions were provided to and considered by the Rhino Pride Foundation, and the final selected solution resulted in camera site locations providing high-quality cover with respect to all the covering objectives, while requiring fewer cameras than initially expected-resulting in significant cost savings and reduced future maintenance and upgrade requirements. The solution approach presented here may be applied to other site-selection problems with similar coverage requirements, including military radar and weapon systems, and wildfire detection systems.

Fire Technology, 2021

Tower-mounted camera-based wildfire detection systems provide an effective means of early forest ... more Tower-mounted camera-based wildfire detection systems provide an effective means of early forest fire detection. Historically, tower sites have been identified by foresters and locals with intimate knowledge of the terrain and without the aid of computational optimisation tools. When moving into vast new territories and without the aid of local knowledge, this process becomes cumbersome and daunting. In such instances, the optimisation of final site layouts may be streamlined if a suitable strategy is employed to limit the candidate sites to landforms which offer superior system visibility. A framework for the exploitation of landforms for these purposes is proposed. The landform classifications at 165 existing tower sites from wildfire detection systems in South Africa, Canada and the USA are analysed using the geomorphon technique, and it is noted that towers are located at or near certain landform types. A metaheuristic and integer linear programming approach is then employed to search 1 for optimal tower sites in a large area currently monitored by the ForestWatch wild-fire detection system, and these sites are then classified according to landforms. The results support the observations made for the existing towers in terms of noteworthy landforms, and the optimisation process is repeated by limiting the candidate sites to selected landforms. This leads to solutions with improved system coverage, achieved within reduced computation times. The presented framework may be replicated for use in similar applications, such as site-selection for military equipment, cellular transmitters, and weather radar.

Frontiers in Ecology and Evolution, 2021

Existing collaborations among public health practitioners, veterinarians, and ecologists do not s... more Existing collaborations among public health practitioners, veterinarians, and ecologists do not sufficiently consider illegal wildlife trade in their surveillance, biosafety, and security (SB&S) efforts even though the risks to health and biodiversity from these threats are significant. We highlight multiple cases to illustrate the risks posed by existing gaps in understanding the intersectionality of the illegal wildlife trade and zoonotic disease transmission. We argue for more integrative science in support of decision-making using the One Health approach. Opportunities abound to apply transdisciplinary science to sustainable wildlife trade policy and programming, such as combining on-the-ground monitoring of health, environmental, and social conditions with an understanding of the operational and spatial dynamics of illicit wildlife trade. We advocate for (1) a surveillance sample management system for enhanced diagnostic efficiency in collaboration with diverse and local partners that can help establish new or link existing surveillance networks, outbreak analysis, and risk mitigation strategies; (2) novel analytical tools and decision support models that can enhance self-directed local livelihoods by addressing monitoring, detection, prevention, interdiction, and remediation; (3) enhanced capacity to promote joint SB&S efforts that can encourage improved human and animal health, timely reporting, emerging disease detection, and outbreak response; and, (4) enhanced monitoring of illicit wildlife trade and supply chains across the heterogeneous context within which they occur. By integrating more diverse scientific disciplines, and their respective scientists with indigenous people and local community insight and risk assessment data, we can help promote a more sustainable and equitable wildlife trade.

World Bank Policy Research Working Papers, 2021

Existing methods of prioritizing rural roads for construction in hilly and mountainous settings r... more Existing methods of prioritizing rural roads for construction in hilly and mountainous settings require expensive data collection or major simplifications of ground conditions. Traditional social surplus based-methods favor economic and political decision criteria over social criteria, despite evidence of the latter's importance, and struggle to scale beyond major roads to feeder roads, forcing local governments with limited capacity to adopt ad-hoc alternative criteria. Using roads proposed for construction in Nepal's remote Karnali province, this paper develops a scalable method to prioritize these roads for inclusion in construction plans with the aim of optimizing potential accessibility improvements to specified services in dry and monsoon seasons – within Karnali's infrastructure budget constraints. Road-specific improvements in accessibility to services are measured by estimating accessibility changes resulting from each proposed road within a multimodal accessibility model. In this paper, walking across Karnali's mountainous, high-elevation terrain is incorporated as a primary modality – a rarity in related accessibility literature. These improvements are implemented within heuristic and integer-linear programming optimization models. Optimization-determined solutions were calculated within a day, and substantially outperformed the actual roads selected by Karnali's provincial government in terms of accessibility, efficiency, and political economy.

International Transactions in Operational research

Effective early detection of forest fires can be achieved by specialised systems of tower-mounted... more Effective early detection of forest fires can be achieved by specialised systems of tower-mounted cameras. Foresters and locals with intimate knowledge of the terrain traditionally plan the tower site locations-without the aid of computational optimisation tools. However, such knowledge and expertise may not be available to system planners when entering vast new territories. The process of selecting multiple tower sites from a large number of potential site locations with the aim of maximising system visibility of smoke above a prescribed region is a complex combinatorial optimisation problem. We present two recent applications of novel site-selection frameworks for tower-mounted camera-based wildfire detection systems (CWDS), which have been under development with guidance from experts from the South African developed ForestWatch wildfire detection system. A novel single-site search framework determined alternatives for 13 proposed sites in South Africa's Mpumalanga province, of which 6 alternatives were chosen over the initially proposed sites. The system site selection framework was showcased in determining a four-camera CWDS layout in South Africa's Southern Cape-significantly improving on the detection capability of the layout initially proposed by technical experts. a 32-m tower on top of which a camera is placed, with the solar power supply visible near the base of the tower (Heyns et al., 2019).

IEEE Access, 2020

Covering location problems aim to find optimal site locations for the placement of systems of fac... more Covering location problems aim to find optimal site locations for the placement of systems of facilities, e.g. cellular transmitters, surveillance sensors, military equipment, and weather radar. The aim is to maximise system coverage over demand regions that are modelled as demand points (targets), which are discretised representations of the terrain surface contained within specified geographical boundaries. Covering problems are computationally difficult to solve and determining globally optimal solutions is typically not possible within realistic computation times. The reduced target strategy is one approach to reduce computational complexity, but preceding research into its applications has been limited to modestly sized study areas, outdated facility specifications, and simplified and impractical modelling approaches without consideration of topography – the practical and computational challenges associated with solving modern facility location problems have been overlooked. A reduced target-resolution strategy is investigated in this paper to solve large, real-world facility location problems with requirements beyond those typically encountered in the literature. Drastic reductions in optimisation computation times are achieved, while improving on the solution quality of previous best efforts. The strategy offers a simple and easy-to-replicate process and does not require any elaborate site/demand abstraction processes or heuristics, and may be beneficial to various modern site-selection problems – particularly in environments in which rapid decision-making is required, and when the problem instance is outside the bounds of tractability for global optimisation.

International Journal of Wildland Fire, 2019

Early forest fire detection can effectively be achieved by systems of specialised tower-mounted c... more Early forest fire detection can effectively be achieved by systems of specialised tower-mounted cameras. With the aim of maximising system visibility of smoke above a prescribed region, the process of selecting multiple tower sites from a large number of potential site locations is a complex combinatorial optimisation problem. Historically, these systems have been planned by foresters and locals with intimate knowledge of the terrain rather than by computational optimisation tools. When entering vast new territories, however, such knowledge and expertise may not be available to system planners. A tower site-selection optimisation framework which may be used in such circumstances is described in this paper. Metaheuristics are used to determine candidate site layouts for an area in the Nelspruit region in South Africa currently monitored by the ForestWatch detection system. Visibility cover superior to that of the existing system in the region is achieved and are obtained in a number of days, while traditional approaches normally require months of speculation and planning. Following the results presented here, the optimisation framework is earmarked for use in future ForestWatch system planning.

Geographical Analysis, 2018

The placement of facilities according to spatial and/or geographic requirements is a popular prob... more The placement of facilities according to spatial and/or geographic requirements is a popular problem within the domain of location science. Objectives that are typically considered in this class of problems include dispersion, median, center, and covering objectives — and are generally defined in terms of distance or service-related criteria. With few exceptions, the existing models in the literature for these problems only accommodate one type of facility. Furthermore, the literature on these problems does not allow for the possibility of multiple placement zones within which facilities may be placed. Due to the unique placement requirements of different facility types — such as suitable terrain that may be considered for placement and specific placement objectives for each facility type — it is expected that different suitable placement zones for each facility type, or groups of facility types, may differ. In this article, we introduce a novel mathematical treatment for multi-type, multi-zone facility location problems. We derive multi-type, multi-zone extensions to the classical integer-linear programming formulations involving dispersion, centering and maximal covering. The complexity of these formulations leads us to follow a heuristic solution approach, for which a novel multi-type, multi-zone variation of the NSGA-II algorithm is proposed and employed to solve practical examples of multi-type, multi-zone facility location problems.

Computers, Environment and Urban Systems, 2016

The placement of certain facilities, such as radars and wind turbines, requires careful planning ... more The placement of certain facilities, such as radars and wind turbines, requires careful planning according to very specific geographical and spatial requirements. Such placement problems are often solved by metaheuristics which find near-optimal solutions within a fraction of the time required to solve these problems exactly. The use of high-resolution representations of the feasible search space generally ensures a high level of solution quality and accuracy, but involves evaluation of a larger number of candidate solutions than lower resolution representations , and is therefore more time-consuming. A trade-off between solution quality and time requirements must therefore be achieved when choosing an appropriate resolution of data to include in geospatial facility location models. In this paper, we propose a novel explore-and-exploit, multi-resolution solution approach that takes advantage of the reduced computational requirements associated with lower resolution terrain data, while simultaneously benefitting from the quality of solutions returned at higher resolutions. Our multi-resolution approach is capable of outperforming analyses in which only highest resolution data are considered, both in terms of solution quality and solution time requirements.

South African Journal of Industrial Engineering, 2015

A recent development within the MeerKAT sub-project of the Square Kilometre Array radio telescope... more A recent development within the MeerKAT sub-project of the Square Kilometre Array radio telescope network was the placement of a network of three observation cameras in pursuit of two specific visibility objectives. In this paper, we evaluate the effectiveness of the locations of MeerKAT observation camera network according to a novel multi-objective Geographic Information Systems-based facility location framework. We find that the configuration chosen and implemented by the MeerKAT decision-makers is of very high quality, although we are able to uncover slightly superior alternative placement configurations. A significant amount of time and effort could, however, have been saved in the process of choosing the appropriate camera sites had our solutions been available to the decision-makers.

MSc Thesis presented in partial fulfilment of the requirements for the degree Master of Science a... more MSc Thesis presented in partial fulfilment of the requirements for the degree Master of Science at the University of Stellenbosch ENGLISH ABSTRACT: In a ground based air defence environment an operator is required to make multiple decisions within a short period of time in order to assign weapons to threatening aircraft. Such assignments are critical to the successful defence of assets on the ground and as a result decision making is extremely stressful. Mathematical modelling of this environment may serve the purpose of aiding the operator in his/her decision making process by providing additional information with respect to various real-time attributes of aircraft and assets. The problem may be divided into two parts, namely that of threat evaluation and that of weapon assignment. The focus in this thesis is on the former subproblem. General requirements for a threat evaluation decision support system are established with respect to the tactical environment, craft and asset attrib...

Other by Andries Heyns

Transportation

Expert Systems With Applications, 2020

Frontiers in Ecology and Evolution, 2020

Journal of Transport Geography, 2021

Computers, Environment & Urban Systems, 2021

Fire Technology, 2021

Frontiers in Ecology and Evolution, 2021

World Bank Policy Research Working Papers, 2021

International Transactions in Operational research

IEEE Access, 2020

International Journal of Wildland Fire, 2019

Geographical Analysis, 2018

Computers, Environment and Urban Systems, 2016

South African Journal of Industrial Engineering, 2015

Applications of the sub-discipline of location science within the larger discipline of operations... more Applications of the sub-discipline of location science within the larger discipline of operations research include the location of important facilities, such as hospitals, fire stations, libraries and depots. Existing facility location resolution techniques are generally based on commercial and transportation modelling criteria with little or no consideration given to the characteristics and influence of surrounding terrain. Recent technological advances have, however, resulted in a subfield of location science rapidly increasing in popularity — the field of geospatial facility location. Geospatial facility location science places a strong emphasis on terrain and environmental factors in the search for suitable sites for facilities with complex location requirements. Examples of such facilities include radars, telecommunication towers, watchtowers and wind turbines.
The applications in which geospatial facility location solution approaches are relevant generally include only one type of facility. As a result, problem-specific solution approaches tailored to the specific facility type considered are usually adopted. On closer inspection, however, the solution approaches followed for these facility location problems reveal striking similarities which may be exploited in a generic manner in order to provide more efficient solution tradeoff alternatives.
By removing the problem-specific approach to facility location modelling and the design of the associated solution techniques, there is an opportunity to develop a generic, dynamically implementable geospatial facility location framework which may be applied to identify facility location trade-off alternatives in various applications. Moreover, networks of multiple facility types may be considered in such a framework — something that is a rarity in the current geospatial facility location literature. In view of the continual advancement of technology, a generic geospatial facility location framework may even prove to be applicable to future facility location problems in which the placement criteria are currently unknown.
The design of a generic geospatial facility location framework is therefore pursued in this dissertation. A dynamic mathematical foundation, which eliminates facility-specific solution methodologies, is established. A computerised concept demonstrator based on this framework is implemented within a generic solution paradigm. This concept demonstrator accommodates as special cases a variety of popular facility location problems from the literature. The geospatial facility location framework is demonstrated to be a dynamic and useful tool for solving complex facility location problems that have not previously been encountered due to the limitations of problem-specific solution search designs.

The purpose of this document is to provide additional information to accompany our paper titled A... more The purpose of this document is to provide additional information to accompany our paper titled A multi-resolution approach towards point-based multi-objective geospatial facility location, which appears in Computers, Environment and Urban Systems. Worked examples of the MRA compared to the SRA for terrain of varying size and complexity, and for problems with diverse criteria included as objectives and constraints are provided in section 1. The search efficiency of our implementation of the NSGA-II is compared to an exact approach in section 2. Pseudocode for the MRA is provided in section 3 and illustrations of the four resolution levels used by the MRA to solve the GFLoP instance in the main document are provided in section 4.

UCGIS Symposium 2020, 2020