Abstracts Track 2026

Area 1 - Methodologies and Technologies
Area 2 - Applications

Area 1 - Methodologies and Technologies

Nr: 87
Title:

Rotational Social Grouping: Algorithms and Classroom Applications
Authors:

Alessandro Hill, Tülay Flam and Steffen Peuker
Abstract: We study new models and methods that can be used to maximally expand the social networks of populations in organized settings over time. Existing ties are complemented by new connections that arise from participation in selected group activities over multiple time periods within a finite horizon. Parameterized by the initial social network, the number of time periods, and the available groups with minimum and maximum staffing requirements in each period, the overall goal is to establish a post-activity social network of maximum density. The underlying complex combinatorial optimization model is a multi-period generalization of the notoriously hard quadratic multi-knapsack problem with a complicated objective function that is of non-linear nature itself. Application areas are broad, spanning organizational contexts, educational environments, social events, and even generic agent-based systems. We develop a series of integer programming formulations including compact linearization, multi-period set partitioning, and novel temporal-star covering. These are solved via branch-and-price algorithms that incorporate problem-tailored complementary column generation techniques, early-termination mechanisms, custom branching strategies, and heuristic improvement methods. We use a diverse set of 800 instances with real social networks, homogeneous and heterogeneous group configurations, and up to 15 rotations to analyze the model and methods. We computationally quantify formulation strengths, preprocessing impact, as well as overall algorithmic performance. We also show that significant performance increases can be obtained from logic-based problem reduction and propagation, leading to tightened group-size bounds and even group elimination. The developed methods are efficient and suitable for practical application with mid-sized populations. We report findings from a comprehensive case study in higher education, involving ten engineering classes over an entire academic term. Using scenarios based on different rotation frequencies, we investigate potential and effectiveness of our method in creating new ties. Our experiments allow us to observe dynamics related to Dunbar’s number. Finally, we perform a qualitative analysis revealing how collaboration time impacts the types of ties that are established.
Nr: 91
Title:

Innovative Approaches to Strategic Decision-Making in Management Science
Authors:

Omar Adel Hamood Aboras and Yao Xun
Abstract: In the face of an increasingly complex and competitive global market, management science is at the forefront of developing advanced strategies that help organizations make more informed, efficient, and impactful decisions. This research delves into innovative decision-making methodologies by combining traditional optimization techniques with emerging technologies, such as machine learning and artificial intelligence. Our study aims to identify novel approaches to strategic decision-making that enhance both operational effectiveness and organizational agility. Central to this research is the integration of data-driven decision models with behavioural theories to bridge the gap between theoretical predictions and practical applications. We examine how companies can optimize resource allocation, improve supply chain management, and better align their strategies with long-term business goals. Through the use of advanced optimization algorithms, such as mixed-integer programming (MIP) and dynamic simulation models, we demonstrate how organizations can solve complex, multi-dimensional problems in areas ranging from financial planning to product development. Additionally, this paper explores the role of machine learning algorithms in predicting future market trends, consumer behaviors, and operational bottlenecks. By analyzing large datasets from diverse sectors, including retail, logistics, and healthcare, we highlight how predictive analytics can empower firms to anticipate challenges and respond proactively, reducing risks and capitalizing on emerging opportunities. The integration of these techniques into real-time decision-making frameworks offers a significant competitive advantage, enabling businesses to make faster, more accurate decisions. The study further investigates the implications of uncertainty in decision-making processes and presents models designed to optimize decisions under conditions of risk and limited information. By analyzing the effects of environmental uncertainty, market volatility, and shifting consumer preferences, we provide actionable insights into how businesses can navigate unpredictable conditions while maintaining operational efficiency and financial sustainability. Through a series of case studies and empirical analyses, we illustrate the practical applications of these strategies across various industries. Our findings show that organizations implementing advanced decision-making frameworks not only achieve better resource utilization but also enhance overall customer satisfaction, drive innovation, and reduce operational costs. This research underscores the importance of integrating modern technological tools with established management principles to create more adaptive, resilient, and forward-thinking organizations. In conclusion, this paper advocates for a paradigm shift in management science where data-driven methodologies and strategic decision-making models work in tandem to shape the future of organizational success. By harnessing the power of these innovative approaches, businesses can not only navigate current challenges but also position themselves for sustained growth in an ever-changing business landscape.
Nr: 110
Title:

An Adaptive Search Framework for Heterogeneous Fleet Routing within ERP/TMS Systems
Authors:

Kahraman Bekir Çetin and A. Serdar Taşan
Abstract: Vehicle planning is becoming more common in ERP and transportation management systems. Constraints constantly evolve, planning data differs by organization, and new business rules must be implemented without repeatedly changing the solver. This study presents an adaptive neighborhood search engine for input data mapping and sampling scenarios in heterogeneous vehicle planning, accounting for real-world constraints (e.g., duty-time policies, eligibility, and capacity limits) and variations. Numerous constraint families can be enabled through configuration. New rules can be added to advanced computational modules at the cost layer level without changing neighborhood and local search operations. Integrated with ERP/TMS, the engine supports modular constraint management tailored to routing objectives. That engine provides evaluable results and allows multi-scale planning within specified runtimes, with calibration based on data size and problem characteristics. The solution engine manages the operator portfolio via adaptive selection, employs local searches for improvement, and operates a controlled acceptance policy to maintain a balanced discovery process within the fixed run-time budget, accounting for available industrial conditions. Self-calibration policies are based on the weights of the constraints and the characteristics of the sample (fleet mix, demand scale, service and time statistics, and distance distribution). It exhibits consistent behavior from small tactical plans to large daily plans. The kernel provides feasible routes with output scores suitable for analysis, which can be managed, supported, and integrated into customizable enterprise decision support systems. Experiments on public benchmarks and ERP-based planning outputs show more efficient solutions and lower total costs in real-world implementations. The program's outputs remain consistent across different aspects, including feasibility rate, component costs, and runtime.
Nr: 113
Title:

To Extend or Not to Extend? Dynamic Shift Lengths in Workforce Planning
Authors:

Alireza Sabouri
Abstract: Emergency Departments face the challenging task of scheduling physicians to meet uncertain patient demand. To better match physician availability with patient arrivals, we consider the possibility of dynamically extending shifts. Our numerical results suggest that shift extensions can reduce expected wait times with the same number of physicians.

Area 2 - Applications

Nr: 83
Title:

Optimized Itinerary Planning for LNG-Powered Cruise Ships
Authors:

Elizabeth Gore and Alessandro Hill
Abstract: The global cruise industry faces increasing pressure to decarbonize in accordance with International Maritime Organization regulations and Net Zero 2050 objectives. Liquified natural gas (LNG) has emerged as a leading transitional fuel, offering significant reductions in carbon dioxide, sulfur oxides, nitrogen oxides, and greenhouse gas emissions compared to traditional marine fuels such as diesel. However, integrating LNG-powered ships introduces operational complexities, particularly in itinerary planning, since limited bunkering infrastructure and fuel range constrain routing flexibility. State-of-the-art planning tools currently lack the capability to account for these challenges and their economic implications. This study introduces novel optimization models for cruise itinerary planning under LNG restrictions. We develop a data-driven decision-support tool for cruise companies that can be used for both strategic and operational planning. The approach is based on an integer programming framework that determines the sequence in which ships visit the set of given ports, as well as arrival and departure times, while minimizing cost and maximizing customer satisfaction. Beyond itinerary-specific requirements, such as port operational hours and nautical limitations, the model incorporates port-dependent refueling options and the vessels’ limited navigational range. This results in a combination of the asymmetric Traveling Salesman Problem and a scheduling problem, that is known to be particularly challenging to solve. In a computational study using a diverse set of current industry data, we investigate how LNG adoption reshapes cruise operations by altering port sequences and refueling patterns. We study a broad set of scenarios, using various ship types and ports across multiple geographical basins, showing that the proposed optimization models can solve realistic problems efficiently. The results reveal trade-offs among key performance indicators, including customer satisfaction, sustainability, and profitability. This research contributes to a broader understanding of technological and operational innovations essential for decarbonizing the global cruise industry. Future work can extend this framework to support tactical planning in related domains such as fleet deployment and port infrastructure development.
Nr: 97
Title:

Economic Analysis and Solution Approaches for Combined Forward-and-Reverse Logistics in Hub-and-Spoke e-Commerce Networks
Authors:

Daniele Manerba, Alessandro Gobbi and Francesca Vocaturo
Abstract: Nowadays, returns management has become an outstanding issue in the e-commerce market, since the underlying operations involve high additional costs and externalities. A well-consolidated strategy for managing e-commerce logistics integrates forward and reverse transportation systems, ensuring the collection of returns alongside traditional product distribution. This approach also employs hub-and-spoke networks to aggregate both distribution and collection demands from several customers into a few central hubs. Within this framework, we study a complex variant of the Vehicle Routing Problem with divisible deliveries and pickups in which each hub may have mandatory delivery and return pickup demands and can be visited multiple times within the same or different routes [1]. Due to the large fluctuation of demand within the aggregating hubs, we also assume that an uncertain optional pickup quantity may arise and that the pickup service for this demand is optional as well. The problem involves two main decisional stages, the day ahead and the operational day times, in which different aspects of the problem must be decided. For this reason, we propose a two-stage Stochastic Programming formulation including ad-hoc recourse actions, namely, the possibility for a vehicle to perform a detour to the depot (where it is possible to unload the already picked up demand) and the opportunity to pay for external spot-market transportation services. To tackle the complexity of solving this model using numerous scenarios, we have developed an exact method based on Integer L-shaped decomposition and ad-hoc valid inequalities. The obtained optimal solutions are used to create an economic analysis of the problem with the aim of providing important managerial insights [2]. Finally, to address very large instances, a Progressive Hedging-based matheuristic approach [3] is also proposed. This method exploits a scenario problem decomposition, an Augmented Lagrangian Relaxation framework, and various heuristic enhancements. Our approach overcomes state-of-the-art solvers in terms of solution quality and efficiency over a representative set of realistic instances and can be readily adapted to similar contexts. 1. Nagy, G., Wassan, N.A., Speranza, M.G., Archetti, C. (2015). The vehicle routing problem with divisible deliveries and pickups, Transportation Science 49 (2), 271–294. 2. Gobbi A., Manerba, D., Vocaturo, F. (forthcoming). Incorporating stochastic optional pickup demand in routing operations with divisible services for hub-and-spoke e-commerce returns management systems (under III round revision on an international journal). 3. Christiansen, J., Dandurand, B., Eberhard, A., Oliveira, F. (2023). A study of progressive hedging for stochastic integer programming, Computational Optimization and Applications 86, 989–1034.