Abstract Artificial Intelligence (AI) has advanced significantly in the 21st century, evolving into a crucial tool for decision-making. A prominent trend is its integration with Multi-Criteria Decision-Making/Aiding (MCDM/A) methods to support complex decisions across diverse engineering domains. This paper presents a systematic literature review analyzing 111 papers from scientific databases on integrating AI with MCDM/A methods. Unlike prior reviews that primarily catalogue methods or hybrid techniques, this study introduces a socio-technical analytical framework comprising three layers—technical configurations, functional mechanisms, and human–AI collaboration patterns—to explain how and why AI reshapes multicriteria decision processes. The findings reveal recurrent architectural patterns, identify dominant functional roles of AI across decision-process phases, and uncover an emerging shift from automation-oriented systems toward augmentation-based decision support. Rather than providing a purely descriptive mapping of the literature, this review undertakes an investigative task guided by a socio-technical framework. By examining how structural configurations of AI–MCDM/A integration reshape the stages of the decision process and redistribute roles between humans and AI, the study moves beyond cataloguing techniques to uncover underlying integration logics, structural tensions, and developmental trajectories. In doing so, it offers both a conceptual consolidation for scholars and a structured foundation for the design of next-generation, human-centered intelligent decision support systems (IDSS).
Abstract This paper analyzes the electricity consumption of Bitcoin mining as a component of blockchain-based financial infrastructure and develops a hybrid forecasting framework that combines a Neural Network Autoregressive model with Exogenous Inputs (NARX) and Mixed Data Sampling (MIDAS). The specification embeds nonlinear state dependence within a feedforward neural network structured as a NARX and exploits mixed-frequency information from daily and monthly indicators to forecast weekly electricity consumption. A key methodological contribution lies in reframing exogenous variable selection as a ranking-based optimization problem grounded in individual explanatory power. To support this, a large language model (LLM)-assisted screening procedure is used to construct a theory-consistent pool of candidate predictors drawn from the finance, energy and cryptocurrency literature. From this pool, an optimization-based strategy identifies a parsimonious subset of variables that minimizes forecast error within the NARX–MIDAS framework. Empirical results demonstrate that the optimized model significantly outperforms benchmark specifications, achieving reductions of 15–20% in root mean squared error and 10–12% in mean absolute error. Beyond predictive performance, the proposed framework yields interpretable insights into how macroeconomic conditions, policy-related uncertainty and financial market dynamics influence Bitcoin mining activity. These findings have direct implications for risk management, energy planning and regulatory oversight in blockchain-based financial systems, highlighting the value of integrating LLM-assisted knowledge extraction with rigorous optimization-driven forecasting methodologies.
Abstract This paper provides a comprehensive historical and methodological review of univariate time-series forecasting from the second half of the 19th century to 2025, with specific emphasis on benchmark methods that have shaped forecasting practice across economics, finance, energy and supply chain management. We trace the evolution from foundational probabilistic theory through adaptive exponential smoothing, the Box–Jenkins ARIMA methodology, non-parametric methods and contemporary hybrid and learning frameworks. Our focus is deliberately on simple, replicable and computationally efficient methods that serve as essential baselines against which more complex approaches must be evaluated. The chronological taxonomy spans eight distinct eras: the pre-1930 foundations; the formative decade (1930–1940); the probabilistic foundations (1940–1949); the empirical shift (1950–1964); the algorithmic and computational expansion (1965–1979); the consolidation phase (1980–1999); the competition and automation era (2000–2019) and the most recent era (2020–2025) introducing hybrid and adaptive learning frameworks. This review is essential, both for tracing the historical roots of forecasting methods and for understanding the basic principles on which benchmark models are built: simplicity, parsimony, computational efficiency, transparency and out-of-sample reliability. This review is addressed to a broad readership: academic researchers and early career scholars tracing the intellectual and methodological history of the field, practitioners seeking reliable and transparent methods for operational forecasting, and management mathematicians and quantitative analysts from cognate disciplines—operational research, decision analytics and management science—for whom benchmark forecasting methods constitute an essential but underexplored toolkit. Benchmark forecasting methods are the foundation of modern predictive analytics.
Abstract The Digital Product Passport (DPP) is increasingly seen as a crucial tool for shifting the construction industry from a linear to a circular economy. However, merely complying with digital DPP requirements may be insufficient to substantially improve physical product circularity. To address this challenge, we develop a three-stage Stackelberg game model for closed-loop supply chains (CLSC) incorporating DPP, aimed at examining DPP decision mechanisms and proposing strategies to encourage DPP adoption. Our findings indicate that large-scale exchange platforms are the most effective reverse channel for collecting used materials. Furthermore, implementing a minimal reuse potential policy for manufacturers, combined with two-part tariff contracts, can enhance CLSC performance. We also propose that circularity inspections of new products can incentivize manufacturers to enhance product circularity. This study provides actionable guidance for companies and policymakers on making retrofit decisions within the DPP framework, with potential benefits including increased product reuse, higher collection rates and broader acceptance of DPP.
We model the reliability of a system with a main component and two auxiliary components for protection, one active, the other on cold standby. The main component performs the system function. The auxiliary components are called protection blocks. The failure rate of the main component depends on whether it is protected or not. The blocks have their own constant failure rate, that is, their lifetimes are exponentially distributed. The reliability of such a system is obtained and the properties of its failure rate are investigated. A necessary condition is obtained to compare the cost-effectiveness of single and two-block protection. Likewise, we obtain necessary conditions for comparing the effectiveness of different three-component designs under the mean time to failure criterion. In this way, the work provides decision support for the designers of systems protected by blocks.
This paper introduces human-centric analytics (HCA) as a design paradigm for integrating mathematical models, data and human judgement into management decision-making. HCA addresses situations in which analytics must be understood, adapted and used within complex organizational settings, rather than treated as standalone technical artefacts. While human-centred design (HCD) offers useful principles for usability and participation, HCA addresses a different design problem. It focuses on the practical design and adaptation of analytical artefacts, techniques and processes, including the negotiation of model assumptions and appropriate levels of granularity. Drawing on a longitudinal case study in a pharmaceutical supply chain, the paper examines a series of analytics interventions involving forecasting, simulation, statistical analysis, visualization and data blending. Analysis of both successful and unsuccessful interventions shows how mathematical tools gained traction when they were developed through iterative engagement with users' work, expertise and constraints. The paper presents HCA as an empirically derived and theoretically grounded design paradigm, supported by an umbrella framework organized around four recurring design activities: structuring perceptions, structuring empirical data, overcoming resistance and evolving solutions. The framework supports flexible combinations of tools and techniques, enabling analytics to become technically grounded, contextually meaningful and integrated into organizational practice. As data-driven systems and AI become increasingly embedded in management, HCA offers an approach for designing analytics that augment human practice rather than bypass it.
How should organizations select board members when they must balance expertise, oversight, diversity and practical constraints such as board size? This paper introduces a structured decision-making framework designed to support board composition under the UK's 'comply or explain' governance regime. The approach first identifies boards that meet minimum requirements for competence, financial oversight, and equity, diversity and inclusion, and then selects those that provide the strongest overall balance of skills while remaining appropriately sized. Two contrasting perspectives are examined. The first, equality of outcome (EOu), incorporates explicit diversity and representation targets. The second, equality of opportunity, focuses on merit and fiduciary capability without imposing outcome-based constraints. Using publicly available data on a major listed company, the analysis shows that boards constructed under EOu can achieve stronger representation while maintaining, and in some cases improving, overall capability. The framework also provides practical diagnostic tools that allow organizations to test whether excluded candidates could reasonably have been selected, offering a transparent way to assess fairness and potential legal or reputational risk. The approach is implemented in a flexible computational setting and can be adapted to a wide range of board and leadership selection problems.
Traditional reliability models often overlook condition monitoring (CM) data, resulting in biased predictions and inefficient maintenance planning. This paper introduces a novel, process-based reliability index, the bivariate-state dependent mean remaining lifetime (BSDMRL), for an $n$-component parallel system operating under a delay-time framework. We develop a heuristic method that not only simplifies the computation of the BSDMRL via a more tractable surrogate index but also reveals key monotonicity properties that enable the design of effective preventive maintenance policies. By embedding this reliability index within a threshold-type preventive maintenance policy and a cost-reward optimization model, we jointly determine optimal inspection intervals and maintenance thresholds to minimize the long-run average maintenance cost. Through numerical experiments and a comparative study, we demonstrate that our reliability-driven approach significantly outperforms models that fail to fully integrate CM information, leading to more robust and cost-effective maintenance decisions.
In this paper, we present a new deterioration-reduction model for performance evaluation and maintenance planning of structural systems. In contrast to existing models, our approach models the joint effect of multiple deterioration phenomena. The underlying deterioration process is the squared norm of an $n$-dimensional Wiener process with drift. Maintenance effects are assumed to be imperfect, modelled through applying the idea of the Arithmetic Reduction of Deterioration $(ARD_{1})$ model. As such, in this paper, to model maintenance decisions, we characterize the deterioration reduction model as a mixture of stochastic motions and a random number of load cycles. Then, using the local features of the characterized model as a decision rule and the average cost rate as a measure of policy, we propose an optimal joint policy of periodic repair and replacement. The latter is determined by the optimal number of imperfect repairs and random load cycles. A numerical example is provided to demonstrate the effectiveness of the proposed approach and examine the behavior of optimal solutions as the model parameters change. It is also demonstrated the proposed model outperforms several models emerging as special cases. Finally, the sensitivity analysis of optimal solutions with respect to three variants of the cost model has been carried out, and the implications are discussed in detail.
Integrating environmental and social imperatives into core business strategy is a fundamental requirement for long-term viability. We analyze the underlying business motivations that should drive corporate action, characterizing the transition toward sustainability as a strategic necessity for survival in a resource-constrained world. We identify major operational levers through which firms can influence their sustainability trajectories and demonstrate how mathematical modelling can help translate high-level ambitions into concrete choices. By conceptualizing sustainability as a continuous process of redesigning operational systems rather than a static state, we clarify the critical interface where operational reality and sustainability goals must meet. Finally, we identify upcoming trends likely to shape the next phase of research and practice in sustainable operations and supply chains.
As competition has shifted from individual firms to entire networks, supply chain management has become central to operations research. The extensive body of related work has advanced both supply chain practice and management mathematics theory. Yet, sustainable supply chain management has not received the same depth of treatment. In this editorial, we argue that sustainability cannot simply be added to existing models through small modifications. It requires developing a brand-new perspective on model-based research. We propose four key features that models for sustainable supply chains must incorporate. First, they need to handle trade-offs across different dimensions of sustainability. Second, they must look beyond single companies to examine whole supply chains. Third, they should account for the wider uncertainties involved in measuring environmental and social impacts. Fourth, they need to embrace circular economy principles. The four articles in this special issue show how new mathematical methods can address these challenges. We conclude by discussing future research directions that would help incorporate the inherent complexity and uncertainty of sustainability problems.
In this paper, we study optimal screening procedures for items with random number of defects. Each defect that causes an item's failure during this screening procedure is repaired/removed. Upon observing the numbers of removed defects, a decision is made whether to discard an item or to justify its future field operation. It is shown that these decisions depend on the distribution of the number of defects in an item. Three discrete distributions are considered: negative binomial, Poisson and binomial. It is shown, e.g. for the negative binomial case, that screening out of items with any number of removed defects improves the quality of remaining items. On the other hand, for the Poisson distribution of defects, there is no need to screen out items, as the distribution of the number of remaining defects does not depend on the number of the removed defects. The optimal screening policies to minimize the corresponding expected cost functions for each case are analyzed. The numerical illustrations of the obtained results are provided. Through the numerical examples, it is shown that the optimal screening policy significantly differs depending on the distribution of the number of defects in an item as well as the involved costs.
An impulsive differential equation model of cancer treatment by radiation therapy (RT) is studied. Analytical results for the model's persistence and eradication of cancer cell volumes are obtained to illuminate the dynamics between tumor growth and RT. It is also shown that, although periodic solutions may exist, they are necessarily unstable. A modified model is then proposed, assuming that RT is more effective than the first model assumes. In addition to similar results as for the original model, conditions are obtained under which periodic solution exists and is globally stable, showing the possibility that regression can occur in periodicity. Numerical simulations are provided to confirm the results.
Managing spare tools and equipment remains a significant operational challenge for aviation Maintenance, Repair and Overhaul (MRO) organizations. For example, despite the 2003 merger of Royal Dutch Airlines (KLM) Engineering & Maintenance and Air France Industries into AFI KLM Engineering & Maintenance, inventory planning and forecasting systems continued to operate separately, leading to parts shortages, excess inventory and extended repair cycles. In several cases, maintenance delays lasted more than a year, indicating that inventory decision-making needs enhancement. This study assesses data-driven inventory management methods to lower costs and improve operational performance in commercial aircraft maintenance. The analysis uses both primary and secondary data to compare traditional inventory-tracking techniques with a more advanced Poisson-based demand-prediction model. The findings suggest that conventional overstocking and understocking strategies increase operational costs and lead to longer maintenance delays, whereas the Poisson model improves forecasting accuracy and cost management. To implement these insights, the study recommends an integrated decision-support dashboard that aligns demand forecasting, inventory policies and standard operating procedures. This dashboard offers MRO managers a practical tool to anticipate demand changes, prevent stockouts, cut holding costs and reduce repair turnaround times. Overall, the study provides actionable insights and a clear roadmap to support the adoption of integrated inventory systems in aviation maintenance operations.
Driven by the concept of low-carbon development, many firms use intelligent technology to promote newly launched green products. Meanwhile, more and more independent remanufacturers (IRs) sell remanufactured products through secondary e-commerce platforms, competing with new products. The promotion of new products will inevitably spill over to remanufactured products, potentially exacerbating the challenge of product cannibalization. However, the existing literature lacks comprehensive insights into firms' green promotion strategies in the presence of secondary platforms and spillover effect. We employ game-theoretic models to integrate promotion strategies for new green product with remanufacturing dynamics, examining their impact on the supply chain and its constituent members. We distinguish between two promotion modes: individual implementation as characterized by non-cooperative strategies employed either by the manufacturer or the retailer, and cooperative strategies based on cost-sharing contracts between the manufacturer and the retailer. Our main findings suggest that the higher promotion effort does not universally translate into increased sales. Only when the manufacturer's promotion capability is relatively limited can the manufacturer and the retailer achieve a win-win situation through cooperation under the scenario where full remanufacturing does not occur. Additionally, despite the potential hindrance of the promotion strategy on IRs' engagement in full remanufacturing activities, consistent promotion proves beneficial in augmenting IRs' profitability. We then contrast our model with the case of Apple Inc, illustrating how promotion spillover causes cannibalization in a firm that combines manufacturing, retailing and remanufacturing functions.
We propose a new method for selecting the local estimation window in forecasting and trading financial returns. The method is built around a particular definition of predictive complexity and we apply it in the simplest of predictors, the sample mean. We derive the exact conditions for the process of optimally selecting the local estimation window among a theoretically found grid of potential values of it. We use different loss functions, statistical and financial, which are first considered individually and then pooled under two selection concepts, stochastic dominance and minimum description length, and find exact expressions as to how the associated complexities and their combinations can be derived and applied. Our results are based on a set of probabilistic assumptions for the time series under study, and, based on those, we offer an inferential procedure for testing the presence of excess trading returns. Our empirical illustration on a set of diverse exchange-traded funds, across different asset classes, suggests that the method works very well in practice and that it can generate both statistical and financial performance enhancements. Extensions to different predictors and different underlying assumptions are discussed.
A value-preserving portfolio strategy is based on the idea that every generation should hand over a portfolio of the same value to the next generation as it received from the foregoing one. Hence, only economic gains can be consumed. Given a natural definition of the portfolio value, the underlying theoretical concept has various features in common with the benchmark approach of Eckhard Platen. In this contribution, we develop a generalization of value preserving portfolio strategies that allows their construction in the framework of the so-called minimal market model of the benchmark approach. This framework is particularly suited for the value-preserving concept as we consider a long-term investment horizon for which the popular constant market coefficient settings based on geometric Brownian motions or jump diffusions are not flexible enough. Further, we demonstrate the application of the value-preserving concept using real market data. In particular, the use of the minimal market model framework allows one to overcome estimation problems for stock returns.
We introduce a robo-advisor system that recommends personalized investment portfolios to users, using a Von Neumann-Morgenstern expected utility model elicited from pairwise comparison data. The robo-advisor system comprises three fundamental components. First, we employ a static preference questionnaire approach to generate pairwise lotteries comparison questions. Next, we design three optimization-based preference elicitation approaches to estimate the nominal utility function pessimistically, optimistically and neutrally. In the preference elicitation process, we assume that the user's pairwise choices are exactly consistent with her/his true utility, without any response error. Finally, we compute portfolios based on the nominal utility using an expected utility maximization model. We conduct a series of numerical tests on a simulated user and some human participants to evaluate the efficiency of the proposed robo-advisor system.
In response to the challenges posed by online-offline competition, many offline sellers have implemented price matching, allowing customers to pay a competitor's lower price for the same product. Meanwhile, online sellers have offered return compensation to offset the inconvenience of product returns. Drawing inspiration from these retail practices, this study develops a duopoly game, where an online seller and an offline seller retail an identical item, to examine the interaction between return compensation and price matching. The findings indicate that, when the number of customers seeking price matching and the travel cost of arriving brick-and-mortar (BM) store are high, the offline seller should conduct price matching regardless of the online seller's return compensation strategy. Conversely, the online seller should only offer return compensation when the offline seller engage in price matching and the travel cost of arriving BM store is significant. However, both price matching and return compensation have the potential to result in a win-win situation for the two sellers. Furthermore, our sensitivity analysis reveals that price matching enables the online seller to derive benefits from offering higher compensation amount.
