Annals of Operations Research

Banerjee, Sougata; Aggarwal, Divya; Sengupta, Pooja

doi: 10.1007/s10479-025-06480-4pmid: N/A

This paper explores the impact of Environmental, Social, and Governance (ESG) related news sentiment and investors' sentiment (IS) on forecasting stock prices by applying machine learning (ML). X (formerly Twitter) data is used to analyze IS using Natural Language Processing (NLP). While ESG sentiment is sourced from the Amenity Analytics dataset, which uses several news sources from LexisNexis to understand the ESG-related sentiment, and both these sentiments are used to forecast stock prices. Several ML models, such as Long Short-Term Memory (LSTM), Random Forest (RF), and Bayesian Ridge Regression (BRR), were employed to predict stock prices for all thirty Dow Jones Industrial Average (DJIA) constituent companies, the top ten companies of S&P 500 ESG Index, and the DJIA index price as a proxy for the US stock market, for the period between January 2018 and December 2020. The results validate the success of the proposed framework. They suggest that adding ESG news sentiment and investors' sentiment to historical stock prices improves the forecast accuracy as measured by MAPE by as much as 1516bps over the base of a 10-day moving average model. Based on the results from the model, we can categorize the stocks into separate groups driven by the predisposition of investors toward the company and ESG news.

Velez, Andres; Wallace, Stein W.

doi: 10.1007/s10479-025-06467-1pmid: N/A

We study the evacuation of large cruise vessels under hazardous conditions. Based on a new sensor mesh technology, allowing wireless two-way communication also when electricity is lost, we propose an optimization-based procedure that provides real-time guidance to passengers. As a first step, we explore a lifeboat assignment procedure to illustrate some of the aspects linked to pedestrian evacuations on vessels and open the discussion for future work using operations research in a context where human behavior is part of the problem.

Dai, Deliang; Javed, Farrukh; Karlsson, Peter; Månsson, Kristofer

doi: 10.1007/s10479-025-06486-ypmid: N/A

Policy institutes such as central banks need accurate forecasts of key measures of economic activity to design stabilization policies that reduce the severity of economic fluctuations. Therefore, this paper develops a kernel ridge regression estimator in a mixed data sampling framework. Kernel ridge regression can handle many predictors with a nonlinear relationship to the target variable. Consequently, it has potential to improve the currently used principal component-based methods when the economic data follow a nonlinear factor structure. In a Monte Carlo study, we show that the kernel ridge regression approach is superior in terms of mean square error and is more robust than principal component-based methods to different nonlinear data generating processes. By using a dataset consisting of 24 economic indicators, we forecast Swedish gross domestic production. The results confirm the superiority of the kernel ridge regression approach. Therefore, we suggest that policy institutes consider the use of kernel-based approaches when forecasting key measures of economic activity.

Maghsoudi, Mehrdad; Noorbakhsh, Alireza; Khanizadeh, Shahrzad; Shokouhyar, Sajjad; Shokoohyar, Sina

doi: 10.1007/s10479-024-06433-3pmid: N/A

Patent data is perhaps one of the most important ways to analyze technology life cycles, as it contains knowledge related to both technological establishments and business viability with objective metrics. The technology lifecycle has four critical phases: emergence, growth, maturity, and saturation. During the phase of emergence, initial patents are somewhat primitive since technologies have just entered the marketplace in their basic forms. Foundational problems and market uncertainties are ironed out during this phase of growth, which then yields to manufacturers’ adoption of the technology and its integration into R&D activities. At the maturity stage, with stiff competition and rapid commercialization, operations gradually render obsolete cutting-edge technology. Technology assumes a foundational role during this maturation phase, open to replacement by newer innovations. The present research covers an analysis of 5,461 patents related to the “Sustainable Supply Chain” through the latent Dirichlet allocation technique, uncovering six different types of technologies with varying lines of time of maturity. One of these technology areas is digital vehicle control, which achieves maturity and paves industries like automotive, aviation, and transportation; it optimizes the flow of sustainable goods with a focus on social and environmental responsibility. While the waste treatment and catalysis sphere are approaching saturation, cost efficiency and eco-conscious material use are in the foreground for sustainable supply chain management. On the other hand, pesticide and pest management are at a presentation phase where extensive R&D verdicts for innovative and eco-friendly product development are going on. This research underlines the role that understanding technology life cycles plays in effective technology portfolio management for market competitiveness. A corporation can, using patent data, assess technology life cycles to identify the best time to invest in R&D versus the time to commercialize technologies. A responsible integration of technology becomes a critical pursuit with looming sustainability goals ahead. This study intends to provide very useful insights into the evolution of sustainable technologies and the complex framework of technology life cycles, in a bid to enable decision-making under highly dynamic circumstances.

Joosten, Reinoud; Harmelink, Rogier; Sparrius, Thom

doi: 10.1007/s10479-024-06462-ypmid: N/A

We model advertising with effects on different time scales for a duopoly in imperfect substitutes using elements from Contest Theory. Firms additionally compete in a short-run strategic variable, here price or quantity, allowing simultaneous or sequential decisions, or collusion in endogenously changing stage games. Strategic variables range from ‘slow’ (advertising), over ‘moderate’ (quantities) to ‘fast’ (prices). We find feasible rewards and equilibria for the limiting average reward criterion. Uniqueness of equilibrium is not guaranteed, and we introduce two criteria which act as natural refinements. We impose stage-game rationality, i.e., the firms play optimally in each stage game. Furthermore, in establishing threats, we require that punishment is sustainable, i.e., the punisher must have nonnegative long term average own profits to avoid bankruptcy.

Chen, Fenge; Peng, Xingchun; Wang, Weile

doi: 10.1007/s10479-025-06488-wpmid: N/A

This paper is devoted to the study of the optimal investment and consumption strategy for an agent who has the addictive habit formation preference. The agent has the constant relative risk aversion (CRRA) preference for consumption and Symmetric Asymptotic Hyperbolic Absolute Risk Aversion (SAHARA) preference towards terminal wealth. Under the criterion of expected utility maximization, the analytical expressions for the optimal investment and consumption strategies are derived by the martingale method and Lagrange dual method. Finally, some sensitivity analyses are presented to illustrate the effects of important parameters on the optimal strategies.

Ferreira, Neuza C. M. Q. F.; Ferreira, João J. M.

doi: 10.1007/s10479-024-06445-zpmid: N/A

The United Nations’ Sustainable Development Goals (SDGs) create a global framework for companies seeking to meet diverse sustainability challenges. Small and medium-sized enterprises’ (SMEs) socioeconomic importance makes their implementation of sustainability initiatives and a strategic orientation toward long-term sustainability quite crucial. This study focuses on creating a multi-criteria decision support analysis system to identify internal initiatives that can improve sustainable entrepreneurship (SE) indices. A mixed-method approach is applied based on cognitive mapping and decision making trial and evaluation laboratory (DEMATEL) technique to cover qualitative and quantitative aspects. The results clarify the causal relationships between six clusters labeled as follows: Technology and Equipment; Collaborative Governance; Social Practices and Community; Environmental Initiatives; Innovation and Entrepreneurship; and Training and Human Capital. Key initiatives that improve SE indices comprise adopting technologically advanced tools, stimulating companies to implement good practices and rewarding good performance, valorizing assets and resources in the regions where companies operate, defining and monitoring environmental goals, enlisting in financial support programs, and concentrating on employees’ quality of life. The findings provide theoretical and practical insights into which strategies significantly promote long-term sustainability. To strengthen the results of the cognitive mapping and DEMATEL analysis, we propose the use of multi-objective programming (MOP) and goal programming (GP) models. These models can help establish objectives that focus on technological advancement, promote best practices, recognize high performance, utilize endogenous resources effectively, define and monitor environmental goals, ensure access to financial support, and enhance employees’ quality of life. Limitations and future research directions are also presented.

Ricciotti, Lorena; Pollice, Alessio

doi: 10.1007/s10479-025-06472-4pmid: N/A

The PublicWorksFinanceIT R package enables users to retrieve and analyze financial data related to public works in Italy. Specifically, it focuses on soil defence investments. The data are sourced from three distinct platforms: the OpenCoesione website, which draws its information from the Cohesion Policy, the OpenBDAP website, the Ministry of Economy and Finance’s open data platform, and the ReNDiS database, provided by ISPRA, which exclusively gathers information about interventions in soil defence. This package offers a user-friendly tool that eliminates the need for direct access to the aforementioned institutional platforms and ensures real-time updates. Additionally, all measurements, metadata, and accompanying analytical tools are provided in English, enhancing accessibility for both international and domestic users. The data records from these three sources are linked using the unique project code (CUP), ensuring that there is no duplication. Moreover, the data is geographically referenced, meaning that each financial investment is associated with a specific municipality within a particular Italian region. This allows to provide information on the region, province, and municipality of each dataset entry. Users can select to geo-reference the data by either the coordinates of the municipality’s centroid or by the polygon representing the municipality’s administrative boundaries. In addition to functions for data retrieval, the package includes functions for visualizing the collected data on maps. After providing a detailed explanation of the purpose and operation of the main commands, the paper presents two case studies illustrating the software’s application. These examples serve as a step-by-step guide to using the PublicWorksFinanceIT package.

Ben Amar, Amine; Féki, Chiheb; Bellalah, Makram

doi: 10.1007/s10479-024-06425-3pmid: N/A

The objective of this paper is to examine the structure as well as the performance of different investment strategies using two asset classes (stocks and commodities) and different portfolio methods. More specifically, we construct different portfolios using three diversification strategies—the traditional minimum-variance portfolio strategy, the minimum-correlation portfolio strategy, and the minimum-connectedness portfolio strategy—and compare them. Our results show that the natural gas market seems to be the most isolated market. Moreover, agricultural commodity markets appear to be broadly insensitive to shocks on non-agricultural commodity markets, which support the thesis stating that commodity markets are potentially segmented. The results also suggest a relatively high connectedness among pure financial markets on average and across the entire sample period, and a relatively low connectedness among regional stock markets and long-term commodity futures markets. The GCC stock market is largely disconnected from other regional financial markets. The portfolio analysis show that the performance of the different investment strategies was largely equivalent in terms of cumulative returns. Furthermore, from mid-2019 to mid-2021, and going through the COVID-19 period, the portfolio strategy composed of regional indices and short-term commodity futures contracts, under the minimum-variance approach, outperforms the other investment strategies. However, the same portfolio strategy, but under the minimum-connectedness approach, outperforms all other investment strategies from mid-2021 and during the ongoing Russian-Ukrainian war period.

Yang, Chih-Hao

doi: 10.1007/s10479-025-06469-zpmid: N/A

In the context of achieving sustainable development goals (SDGs), digital energy planning has become increasingly significant and complex due to the need to consider multiple indices. This complexity poses challenges for decision-makers who must integrate both qualitative and quantitative approaches, particularly in relation to digital energy business models. This study identifies key indicators for decision-making regarding digital energy business models and prioritizes alternatives from environmental, social, governance (ESG), and financial perspectives. Additionally, it incorporates activity-based costing techniques and resource constraints into the multi-criteria decision-making (MCDM) approach. The study employs the decision-making trial and evaluation Laboratory (DEMATEL), the analytic network process (ANP), the VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR), and zero–one goal programming (ZOGP) to optimize digital energy business model portfolios. The findings indicate that perspectives on environmental concern and governance practices significantly influence other dimensions. Furthermore, managing carbon footprint is crucial for evaluating digital energy business models. Specifically, optimal portfolios of digital energy business models—such as green energy experience communities, green energy resource-sharing platforms, asset activation innovative designs, and series connections with grid data demonstrate feasibility and reasonability in terms of resource requirement. This study offers insights into various MCDM approaches, emphasizing the importance of digital energy resource planning over traditional decision-making processes. It provides valuable scientific guidance for optimizing digital energy business model planning and contributes to the transformation of Taiwan's energy industry toward sustainable development.