Annals of Operations Research最新文献

Both the extreme shock model and the (delta )-shock model have been extensively studied and applied in various fields. Nevertheless, the research of these shock models under discrete-time conditions remains limited. Under discrete-time conditions, this paper develops a multi-state mixed (delta )-shock model that incorporates both a generalized (delta )-shock and an extreme shock. The research has the following two main features: on the one hand, the arrival process of the shocks is generalized by assuming that the shock intervals follow a discrete PH distribution; on the other hand, by constructing a discrete-time Markov chain, the discrete PH distribution representations of the lifetime, operating time in perfect state and operating time in inferior state of the system are given, and the expected values and probability mass functions of the corresponding measures of the system are derived through these discrete PH distributions. Additionally, by means of the numerical experiment, the relationship between various measures and parameters of the system is elucidated. The proposed mixed (delta )-shock model can degenerate into the extreme shock model or the (delta )-shock model under certain conditions, thereby providing a new perspective and approach for comprehending and utilizing these shock models.

This study developed an approach to determine the staffing needs of administrative, professional, and technical personnel that does not rely on subjective input. Our method involves a detailed description of work processes and a time study using a web application similar to a timesheet. We determine staffing needs by assessing the workload for each task and calculating the required staffing level based on the total workload. The time study revealed an uneven distribution of workload across tasks and an unbalanced allocation based on the frequency of task performance. It also showed a positive relationship between task execution frequency and workload. Based on these findings and other data trends, we developed task workload predictors and trained a generalized regression model using time study data from various industries. Staffing needs are compared in two ways: (i) using a machine-learning model instead of expert estimates, and (ii) using a bottom-up approach that incorporates time study data and employee feedback. Results indicate that staffing levels derived from the machine-learning model are similar but more conservative than those obtained through the integrated approach, which includes time study data and employee feedback.

Since the creation of the Portuguese National Health Service (SNS, from the Portuguese abbreviation of Serviço Nacional de Saúde), it is the State’s responsibility to ensure the access of the population to quality healthcare, regardless of its economic and social conditions. Efforts for attaining efficiency and reducing wastage, brought along by the successive reforms of the SNS, continue, as the 2009 financial crisis still lingers its effects on the social performance of the provided health services (now again affected by the COVID-19 pandemic). Public hospitals in Portugal are paid following a prospective budget scheme. These contracts define the level of production for each hospital through the computation of unitary costs but do not include variables concerning the access to and quality of health services, which, ultimately, may lead to inefficiencies and misallocation of resources. This study proposes a framework that considers the quality, access, efficiency, and complexity of the Portuguese public hospitals in their payments through the optimisation of the unitary costs established in the contracts, by testing and adapting a ‘best practices payment-based’ tool. Significant cost savings of 64 million €  arose for the inpatient service of a relatively small sample of nineteen hospitals, showing high potential for cost containment while aiming for quality, access, and efficiency improvements.

Emerging and re-emerging infectious diseases continue to be a major threat to human health, and a threat to economic and financial stability worldwide. Epidemiological models have proven to be important tools for informing decisions on optimal policies, interventions, and control measures. However, the importance of surveillance and early detection is often overlooked, and the focus on control measures is mostly in terms of existing infections. In this work we present a framework for deciding on the optimal surveillance scheme that will minimize the overall cost of a disease, balancing the costs pertaining to surveys for early detection with other disease-related costs prior to detection. We show that the expected cost can be fully described for a general case of a stationary policy, thus supporting informed decision making between a few (fixed) strategies. In addition, we show that for time-dependent monitoring policies, the optimal policy is a bang-bang policy, whereby resources should be utilized for monitoring only after a critical time-point, to the maximum extent possible.

Fueled by advances in both robust optimization theory and reinforcement learning (RL), robust Markov Decision Processes (RMDPs) have garnered increasing attention due to their powerful capability for sequential decision-making under uncertainty. In this paper, we provide a comprehensive overview of the theoretical foundations and recent developments in RMDPs, with a particular emphasis on ambiguity modeling. We examine the "rectangular assumption", a key condition ensuring computational tractability in RMDPs but often resulting in overly conservative policies. Three widely used rectangular forms are summarized, and a novel proof is provided for the NP-hardness of non-rectangular RMDPs. We categorize RMDP formulation approaches into parametric, moment-based, and discrepancy-based models, analyzing the trade-offs associated with each representation. Beyond the traditional scope of RMDPs, we also explore recent efforts to relax rectangular assumptions and highlight emerging trends within the RMDP research community. These developments contribute to more practical and flexible modeling frameworks, complementing the classical RMDP results. Relaxing rectangular assumptions tailored to operations management is a promising area for future research, and there are also opportunities for further advances in developing fast algorithms and provably robust RL algorithms.

Initial Coin Offerings (ICOs), known for their regulatory challenges, frequently emerge as hotspots for fraudulent activities, thereby causing substantial losses for investors. Nevertheless, discerning whether an ICO is implicated in scams is exceptionally challenging. To this end, applying machine learning techniques to identify fraud in ICOs has become a multifaceted endeavor. First, we innovatively incorporated multi-classifier integration into a two-stage Positive-Unlabeled (PU) learning framework based on traditional PU learning. This approach employs cross-validation to dynamically calibrate the confidence interval of positive samples and optimizes the mechanism for filtering negative samples, thereby accurately identifying scams in ICO projects. Second, we evaluated the performance of various machine learning algorithms in detecting ICO scams, with CatBoost emerging as the most accurate model. Third, using the SHAP theory, we further analyzed the contribution of constructed features to predicting the risk of ICO fraud, revealing the pathways through which key influencing factors operate. Our findings reveal that PU learning is an effective method for labeling ICO project scams. And it is worth noting that the presence of venture capital (VC) support markedly reduces the likelihood of scams, which underscores the importance in mitigating risk. This study not only enhances the transparency of data-driven decision-making in FinTech but also provides valuable insights for ICO project initiators, investors, and regulatory bodies in their efforts to combat fraudulent activities.

Ensuring resilience and transparency in supply chain (SC) operations has become crucial in today's dynamic and complicated global business environment, particularly in the Fast-Moving Consumer Goods (FMCG) industry. Disruptions like the COVID-19 pandemic have highlighted the weaknesses of SCs, stressing the necessity for proactive initiatives to improve adaptation and visibility. By defining the key success factors (CSFs) that support resilience and transparency in SCs and investigating the role of developing digital technologies, most specifically AI, in accomplishing these goals, this study seeks to solve these issues. This study systematically finds, ranks, and assesses CSFs and their compatibility with cutting-edge technologies using a hybrid methodology that combines DEMATEL, the Best–Worst Method (BWM), and the VIKOR technique. This study highlights AI's unique potential to promote moral decision-making, increase accountability, and improve predictive and adaptive skills. It also presents AI as a transformational facilitator. AI is distinguished from conventional technology by these qualities, which also establish it as a fundamental component of transparent, robust, and sustainable SCs. This study fills a significant gap in the literature by integrating ethical AI holistically within a larger framework of digital technology. By connecting theoretical understanding with real-world implementations, the study adds to the growing conversation around SC transparency and resilience, especially in the FMCG sector. Important results show that using AI gives SCs a competitive edge in negotiating ambiguities and disruptions by allowing them to strike a compromise between operational effectiveness and moral concerns. Through the creative application of cutting-edge digital technologies, the study offers practical insights for both practitioners and scholars, laying the groundwork for more robust and transparent SC ecosystems.

Reducing carbon emissions in the manufacturing sector is an important pathway to achieving sustainable development goals (SDGs). Anchored in SDG 13, this study focuses on the operational and low-carbon research and development (R&D) issues faced by firms in a supply chain, comprising one supplier and one manufacturer, in two scenarios: cooperative decision making and independent decision making. The independent decision-making scenario is classified into three modes: supplier-independent R&D, manufacturer-independent R&D, and two firms simultaneously R&D. First, this study evaluates the various modes with regard to unit abatement levels, corporate profits, environmental impact, and consumer surplus. Next, the analysis investigates the influence of carbon tax and R&D cost coefficient on equilibrium strategy. Our findings indicate that under the cooperative mode, when the cost coefficient is below a certain threshold, a high carbon tax can effectively incentivize both firms to increase their carbon reduction efforts, thereby enhancing the overall profitability of the supply chain and advancing the achievement of the SDG 13. The simultaneous independent R&D can enable all parties to obtain a high unit abatement level than when a single firm conducts independent R&D. In addition, the simultaneous R&D mode of two firms can generate higher profits and consumer surplus than the R&D mode of a single firm, and can bring higher total supply chain profits than the cooperative mode when the R&D cost is low. After extending modes to the case where the two firms cooperate only in the field of R&D, this study finds that the unit abatement level reaches the maximum.

Competitive balance, which refers to the level of control teams have over a sports competition, is a crucial indicator for tournament organisers. According to previous studies, competitive balance has significantly declined in the UEFA Champions League group stage over the recent decades. Our paper introduces alternative indices to investigate this issue. Two ex ante measures are based on Elo ratings, and four dynamic concentration indicators compare the final group ranking to reasonable benchmarks. Using these indices, we find no evidence of any long-run trend in the competitive balance of the UEFA Champions League group stage between the 2003/04 and 2023/24 seasons.

This paper deals with a Poisson input infinite-buffer single-server queue, where the arrivals occur in singles and the server serves the customers in batches. The server serves customers in batches of maximum size “b” with a minimum threshold size “a”. The service time of each batch follows general distribution (including heavy-tailed distribution) independent of each other as well as of the arrival process. The probability generating function (pgf) of the queue-length distributions at an arbitrary epoch as well as at a post-departure epoch of a batch have been derived using the embedded Markov chain and the argument of the rate-in and rate-out principle. The Laplace-Stieltjes transform (LST) of the actual waiting-time distribution (in the queue) of a random customer has also been derived using functional relation between pgf’s. The proposed analysis is based on the roots of the characteristic equation associated with the LST of the waiting-time distribution (in the queue) of a random customer. Using LSTs, the closed-form expressions for the probability density functions and for an arbitrary number of moments of the waiting-time distributions have been presented. We have also done numerical implementation of this procedure for the case of a bulk service infinite-buffer queueing model, and obtained the probability density function for waiting-time distribution of a random customer in the queue.