Integrating machine learning and operations research methods for scheduling problems: a bibliometric analysis and literature review

Fast screening of COVID-19 inpatient samples by integrating machine learning and label-free SERS methods.

Contributors

This study explores how Machine Learning (ML) is being applied across seaport–hinterland networks to improve operational decision-making in container shipping. It draws on a comprehensive systematic review of research published from 2008 to 2025, complemented by a bibliometric analysis of publication trends. The reviewed literature is organised by ML methods, the operational challenges they target, and the data required for model development. Among the various applications identified, container throughput forecasting is the most common. Frequently used datasets include equipment counts, container throughput records, quay crane performance data, truck traffic volumes, and weather information. Findings reveal a clear shift toward combining ML with hybrid frameworks, operations research techniques, and simulation-based models, resulting in stronger prediction accuracy and more robust decision support. The analysis also illustrates the ways ML contributes to decision-making across seaport-hinterland operations, outlining emerging research avenues, stakeholder-driven trends, key implementation issues, and practical recommendations. Overall, the study provides a structured synthesis of current knowledge, mapping major themes and developments while offering methodological insights for future ML research in this domain.

This review paper provides an in-depth exploration of optimization techniques in Operations Research (OR), highlighting their significance, recent advancements, challenges, and future directions. Operations Research utilizes mathematical and analytical methods to optimize decision-making processes across various industries. The paper begins with an overview of OR, discussing its definition, scope, and historical development. It then delves into the fundamentals of optimization, covering different types of optimization problems and algorithms. Classical optimization techniques such as Linear Programming, Integer Programming, and Nonlinear Programming are discussed, followed by an exploration of heuristic and metaheuristic optimization techniques including Genetic Algorithms, Simulated Annealing, Tabu Search, Particle Swarm Optimization, and Ant Colony Optimization. Recent advances in optimization, including hybrid methods, multi-objective optimization, optimization for big data, and integration with machine learning and artificial intelligence, are analyzed. Furthermore, the paper examines the challenges faced by optimization research and outlines emerging trends and future prospects. By synthesizing current knowledge and identifying areas for future research, this paper aims to contribute to the ongoing development and application of optimization techniques in OR and related fields.

Artificial intelligence has been applied in wildfire science and management since the 1990s, with early applications including neural networks and expert systems. Since then, the field has rapidly progressed congruently with the wide adoption of machine learning (ML) methods in the environmental sciences. Here, we present a scoping review of ML applications in wildfire science and management. Our overall objective is to improve awareness of ML methods among wildfire researchers and managers, as well as illustrate the diverse and challenging range of problems in wildfire science available to ML data scientists. To that end, we first present an overview of popular ML approaches used in wildfire science to date and then review the use of ML in wildfire science as broadly categorized into six problem domains, including (i) fuels characterization, fire detection, and mapping; (ii) fire weather and climate change; (iii) fire occurrence, susceptibility, and risk; (iv) fire behavior prediction; (v) fire effects; and (vi) fire management. Furthermore, we discuss the advantages and limitations of various ML approaches relating to data size, computational requirements, generalizability, and interpretability, as well as identify opportunities for future advances in the science and management of wildfires within a data science context. In total, to the end of 2019, we identified 300 relevant publications in which the most frequently used ML methods across problem domains included random forests, MaxEnt, artificial neural networks, decision trees, support vector machines, and genetic algorithms. As such, there exists opportunities to apply more current ML methods — including deep learning and agent-based learning — in the wildfire sciences, especially in instances involving very large multivariate datasets. We must recognize, however, that despite the ability of ML models to learn on their own, expertise in wildfire science is necessary to ensure realistic modelling of fire processes across multiple scales, while the complexity of some ML methods such as deep learning requires a dedicated and sophisticated knowledge of their application. Finally, we stress that the wildfire research and management communities play an active role in providing relevant, high-quality, and freely available wildfire data for use by practitioners of ML methods.

Contributors

Chemical heuristics have been fundamental to the advancement of chemistry and materials science. These heuristics are typically established by scientists using knowledge and creativity to extract patterns from limited datasets. Machine learning offers opportunities to perfect this approach using computers and larger datasets. Here, we discuss the relationships between traditional heuristics and machine learning approaches. We show how traditional rules can be challenged by large-scale statistical assessment and how traditional concepts commonly used as features are feeding the machine learning techniques. We stress the waste involved in relearning chemical rules and the challenges in terms of data size requirements for purely data-driven approaches. Our view is that heuristic and machine learning approaches are at their best when they work together.

In this golden age of artificial intelligence, transparency and responsible decision-making are paramount. While machine learning (ML) and operational research (OR) optimisations are fundamental aspects of AI, the benefits of explainable AI (XAI) for combinatorial optimisations remain underexplored. This study investigates the convergence of XAI and OR, emphasising the importance of transparency in combinatorial optimisations. Using the Knapsack problem as an example, we demonstrate that interpretable ML models can effectively solve combinatorial optimisation challenges and enhance transparency. Additionally, we illustrate the application of post-hoc XAI methods to OR optimisations solved with ML, providing transparent, human-friendly explanations. The key contributions of this work include proposing the application of the SAGE framework for transparent OR, demonstrating the integration of XAI with combinatorial optimisations, and offering practical guidelines for creating transparent explanations. These contributions can aid decision-makers in understanding, communicating, and trusting combinatorial optimisation solutions, paving the way for enhanced transparency in operational research across various sectors.

Innovative Operational Research practitioner, intellectual leader in the development of systems ideas, and founder of the field of management cybernetics. Born: September 25, 1926 in London, England. Died: August 23, 2002 in Toronto, Canada. Education: D.Sc., University of Sunderland, 2000. Key positions: Production Controller and Head of OR Group, Samuel Fox (1949–1956); Head, Operational Research and Cybernetics, United Steel (1956–1961); Managing Director, SIGMA (1961–1966); Development Director, International Publishing Corporation (1966– 1970); Consultant (1970–2002). Awards and Recognition: Silver Medal, Royal Swedish Academy for Engineering Sciences (1958); Lanchester Prize, Operations Research Society of America (1966); McCullough Plaque of the American Society for Cybernetics (1970); Norbert Weiner Gold Medal, World Organization of Systems and Cybernetics (1984); Lifetime Achievement Award, UK Systems Society (1999); Honorary degrees: Doctor of Laws, Concordia University, Montreal, Canada (1988); Doctor of Economic Sciences, University of St. Gallen, Switzerland (2000). Key OR roles: President, Operational Research Society (1970–1971); President, Society for General Systems Research (1971–1972); President, World Organization of Systems and Cybernetics (1970–the time of his death). Stafford Beer was one of the most remarkable figures that Operational Research (OR) in any country has produced. A charismatic, even flamboyant, character, he founded two major and pioneering OR groups; wrote some of the best books about the subject; and was a world leader in the development of systems ideas. He is widely acknowledged as the founder of the field of International Transactions in Operational Resarch.

Machine learning (ML) methods for the analysis of electrocardiography (ECG) data are gaining importance, substantially supported by the release of large public datasets. However, these current datasets miss important derived descriptors such as ECG features that have been devised in the past hundred years and still form the basis of most automatic ECG analysis algorithms and are critical for cardiologists’ decision processes. ECG features are available from sophisticated commercial software but are not accessible to the general public. To alleviate this issue, we add ECG features from two leading commercial algorithms and an open-source implementation supplemented by a set of automatic diagnostic statements from a commercial ECG analysis software in preprocessed format. This allows the comparison of ML models trained on clinically versus automatically generated label sets. We provide an extensive technical validation of features and diagnostic statements for ML applications. We believe this release crucially enhances the usability of the PTB-XL dataset as a reference dataset for ML methods in the context of ECG data.

Basically, Data Mining (DM) and Operations Research (OR) are two paradigms independent of each other. OR aims at optimal solutions of decision problems with respect to a given goal. DM is concerned with secondary analysis of large amounts of data (Hand et al., 2001). However, there are some commonalities. Both paradigms are application focused (Wu et al., 2003; White, 1991). Many Data Mining approaches are within traditional OR domains like logistics, manufacturing, health care or finance. Further, both DM and OR are multidisciplinary. Since its origins, OR has been relying on fields such as mathematics, statistics, economics and computer science. In DM, most of the current textbooks show a strong bias towards one of its founding disciplines, like database management, machine learning or statistics. Being multidisciplinary and application focused, it seems to be a natural step for both paradigms to gain synergies from integration. Thus, recently an increasing number of publications of successful approaches at the intersection of DM and OR can be observed. On the one hand, efficiency of the DM process is increased by use of advanced optimization models and methods originating from OR. On the other hand, effectiveness of decision making is increased by augmentation of traditional OR approaches with DM results. Meisel and Mattfeld (in press) provide a detailed discussion of the synergies of DM and OR.

The paper describes applications of Operations Research (OR) theory and techniques used to solve various types of social problems occurring in our social system. Social systems analysis has for quite some time been the main analytical and scientific approach used to investigate systems and to solve various problems related to modern social systems, including industry, business, the military, public administration, politics, and society in general. We will present here three major roles that operations research (OR) and social systems analysis (SSA) technique have played both practically and theoretically in the solution of social systems problems since it was developed almost 60 years ago. Firstly, we explain briefly OR, SSA, and public policy (PP) regarding what they are, how OR can be contributing to SSA and PP, and how traditional academic disciplines are related each other with the SSA. Secondly, we introduce several examples of the quantitative data analysis, which we have investigated in our school (National Graduate Institute for Policy Studies) to solve various types of social problems including population, traffic and accident, higher education policy, energy policy, and agriculture policy data analyses. Thirdly, we give mathematical modeling analysis with its application to the optimal location model analysis for integrating promotion branch offices in the local government. Fourthly, as an important role of OR as a theory building analysis technique, we explain two problems of apportionment problem and shortest path counting problem. Finally, in the summary section future perspectives of OR are given.

Computer-integrated manufacturing is a notable feature of Industry 4.0. Integrating machine learning (ML) into edge intelligent Industrial Internet of Things (IIoT) is a key enabling technology to achieve intelligent IIoT. To realize novel intelligent applications of edge-enhanced IIoT, ML methods are proposed to improve the cognitive ability of edge intelligent IIoT in this article. First, an ML-enabled framework of the cognitive IIoT is proposed. Second, the ML methods are presented to enhance the cognitive ability of IIoT including the ML model of IIoT, data-driven learning and reasoning, and coordination with cognitive methods. Finally, with a focus on the reconfigurable production line, a scenario-aware dynamic adaptive planning (DAP) with deep reinforcement learning (DRL) was conducted. The experimental results show that the DRL-based dynamic adaptive planning (DRL-based DAP) had good performance in an observable IIoT environment. The main purpose of this work is to point out the effects of ML-based optimization methods on the analysis of industrial IoT from the macroscopic view.

The potential for operations research (OR) with farmer-supplied data coupled with machine learning (ML) to improve crop management is explored through a series of case studies from developing countries. The information provided by the farmers ranged from solely yield to a description of the management of the crop and some details of the growth environment. The climate or weather conditions of the georeferenced farms were estimated from publicly available databases. Two principal analytical approaches were used. The first benchmarks crop performance against farmers' practices and the second establishes relatively homogenous environmental conditions (HECs) in which the variation in crop response is due to variation in management practices and not to spatiotemporal variation in biophysical factors. Both approaches depend on large amounts of data that can only realistically be obtained from records of on-farm experiences using an OR focus. ML effectively defined HECs for crops with limited prior knowledge on the biophysical factors that influence crop response. The definition of HECs facilitated the identification of either individual farmers who managed their crops well within individual HECs or combinations of management practices well suited to the specific spatiotemporal environmental conditions. This opens the way for farmers to learn better agricultural practices from others in the same HEC. Variation in yield and fertilizer response was associated with variation in the El Niño Southern Oscillation (ENSO) patterns up to 24 months before the harvest: this offers the opportunity for farmers to minimize risk, based on ENSO predictions, even when they have no information on how ENSO influences their weather patterns. Despite concerns about the quality of farmer data, the consistency of the analyses suggests that even relatively crude production data from individual farms analyzed with ML can provide useful guidelines for crop management. Limited variation in management on farmers’ fields may limit the ability to identify optimal practices, however, this constraint can be partially obviated by superimposing varied management practices on farmers’ fields. The use of OR combined with ML complements, rather than replaces, traditional research methodologies. Furthermore, the approach must be used carefully with emphasis on the dangers of extrapolation to circumstances that are not encompassed by the original datasets.

A dose deposition matrix (DDM) prediction method using several voxel features and a machine learning (ML) approach is proposed for plan optimization in radiation therapy. Head and lung cases with the inhomogeneous medium are used as training and testing data. The prediction model is a cascade forward backprop neural network where the input is the features of the voxel, including 1) voxel to body surface distance along the beamlet axis, 2) voxel to beamlet axis distance, 3) voxel density, 4) heterogeneity corrected voxel to body surface distance, 5) heterogeneity corrected voxel to beamlet axis, and (6) the dose of voxel obtained from the pencil beam (PB) algorithm. The output is the predicted voxel dose corresponding to a beamlet. The predicted DDM was used for plan optimization (ML method) and compared with the dose of MC-based plan optimization (MC method) and the dose of pencil beam-based plan optimization (PB method). The mean absolute error (MAE) value was calculated for full volume relative to the dose of the MC method to evaluate the overall dose performance of the final plan. For patient with head tumor, the ML method achieves MAE value 0.49 × 10-4 and PB has MAE 1.86 × 10-4. For patient with lung tumor, the ML method has MAE 1.42 × 10-4 and PB has MAE 3.72 × 10-4. The maximum percentage difference in PTV dose coverage (D98) between ML and MC methods is no more than 1.2% for patient with head tumor, while the difference is larger than 10% using the PB method. For patient with lung tumor, the maximum percentage difference in PTV dose coverage (D98) between ML and MC methods is no more than 2.1%, while the difference is larger than 16% using the PB method. In this work, a reliable DDM prediction method is established for plan optimization by applying several voxel features and the ML approach. The results show that the ML method based on voxel features can obtain plans comparable to the MC method and is better than the PB method in achieving accurate dose to the patient, which is helpful for rapid plan optimization and accurate dose calculation. Establishment of a new machine learning method based on the relationship between the voxel and beamlet features for dose deposition matrix prediction in radiation therapy.