2025_004

Drones improve the efficiency and speed of response and reduce the risk to human lives by performing tasks in hazardous or inaccessible areas. Their flexibility in deployment makes them a powerful tool in disaster relief and recovery operations.

Governments often implement lockdowns and enforce social distancing measures during epidemics, particularly when vaccines and cures are not readily available. In such scenarios, drones can play a crucial role in minimizing person-to-person contact by taking over tasks traditionally carried out by humans, such as delivering essential supplies, collecting medical samples, and monitoring body temperatures.

Beyond these tasks, drones have a wide range of other applications, including surveillance of quarantined areas, disinfecting public spaces, and assisting with the transportation of medical equipment to remote or inaccessible regions. Drones can also be used for real-time data collection, helping authorities track and respond to the spread of the virus more effectively.

This topic explores a new approach to humanitarian logistic problems with the application of multiple drones for disaster relief. Additionally, it studies alternatives of truck-and-drone combinations to minimize delivery time and costs, along with a sensitivity analysis.

2025_005

In the aftermath of a disaster, drones can be invaluable for a wide range of tasks due to their speed, flexibility, and ability to access hard-to-reach areas.

Drones can be involved in Search and Rescue Operations, effectively locating survivors and assessing damage. Drones equipped with thermal imaging cameras can scan large areas quickly to detect heat signatures of survivors trapped under rubble, in forests, or in floodwaters, and provide real-time aerial footage to help rescue teams assess the extent of damage and prioritize areas for immediate intervention.

In areas cut off from traditional transportation routes due to collapsed bridges, landslides, or flooding, drones can deliver critical medical supplies such as vaccines, first-aid kits, or blood units. Drones can deliver food, water, and other essential supplies to remote or isolated locations where ground-based delivery is impossible or delayed.

Drones equipped with high-resolution cameras can fly over disaster-stricken areas to survey damage to infrastructure. This allows governments and aid agencies to prioritize repair work and mobilize resources efficiently. Drones can inspect power lines, gas pipelines, and communication towers to assess damage, enabling quicker restoration of essential services.

Drones can be utilized to create up-to-date 3D maps of affected areas, providing detailed information for disaster response teams. This helps responders understand the landscape changes and plan their operations more effectively. They can gather data on environmental hazards such as landslides, flash floods, or chemical spills, aiding in evacuation plans or identifying further risks.

As well as drones can provide recording for legal and insurance purposes. Footage from drones can serve as documentation for legal claims, insurance assessments, and post-disaster evaluations.

This topic explores the potential of using drones in emergency response operations. As a solution for a real-life problem based on a realistic case study, it proposes an efficient emergency response plan with the application of drones.

2025_006

Several logistical challenges can be modelled as pure binary integer programming (BIP) problems. Examples include facility location problems, cutting stock problems, facility location problems, and airline crew scheduling, in addition to many other planning problems. The resulting problems are hard to solve, and heuristic solution methods are often developed for the particular problem class at hand, rather than addressing the general BIP.

However, some heuristics for the BIP have been implemented and tested as a part of two master theses at Molde University college [1, 2] and in related research [3, 4].

One challenge when using heuristics for BIP is the lack of information about solution quality, as only primal bounds are sought. In this thesis topic we propose to investigate a new idea for generating dual bounds: using decision diagrams for optimization.

The core of the idea is to use a form of dynamic programming on a relaxation of the problem [5]. This allows us to calculate an upper bound (for a maximization problem). One may also consider restricted versions of the problem and use that to calculate lower bounds (for a maximization problem). The proposed research question is: how good are the upper bounds that can be found in reasonable time using relaxed decision diagrams for BIPs?

This topic is suitable for one or more students with high ambitions and good programming skills. It is possible to work within an existing framework for solving BIPs implemented in C++.

Contacts: Lars Magnus Hvattum and Bård Inge Pettersen

[1] A. Reznik. 2021. Heuristics for binary integer programming problems. Master thesis, Molde University College, https://himolde.brage.unit.no/himolde-xmlui/handle/11250/2779735

[2] M .Drozd. 2024. Adaptive large neighborhood search for binary integer programming problems. Master thesis, Molde University College, https://himolde.brage.unit.no/himolde-xmlui/handle/11250/3196905

[3] H. Bentsen, A. Hoff, and L.M. Hvattum. Exponential extrapolation memory for tabu search. EURO Journal on Computational Optimization, 10, 100028, 2022.

[4] K. Danielsen and L.M. Hvattum. 2025. Solution-based versus attribute-based tabu search for binary integer programming. International Transactions in Operations Research, forthcoming.

[5] W.J. van Hoeve. 2024. An Introduction to Decision Diagrams for Optimization. TUTORIALS in Operations Research.

2025_018

Background:
Increased urbanization has transformed freight and passenger transportation within urban areas into a critical challenge. Urban transportation is a lifeline for citizens, urban retailers, and industries. At the same time, it has a significant negative impact on the quality of life in cities through congestion, emissions, and space consumption. Urban logistics initiatives have long emphasized the need for collaborative and environmentally friendly solutions to mitigate these effects. However, they face organizational and technological barriers that weaken their long-term robustness.

Recent research suggests that sharing economy initiatives have been presented as potential solutions for greener and more affordable transport (Cassetta et al., 2017; Vazifeh et al., 2018). However, their role in strengthening resilience – the capacity of urban transport systems to adapt, recover, and maintain critical functions under stress – remains underexplored.

Research focus:
This thesis will explore how sharing economy principles can contribute to resilience in urban passenger transport, with a particular focus on the social dimension of sustainability.

Methodology:
The thesis is expected to be qualitative in nature, based on one or more case studies. Candidates are encouraged to collect empirical data through interviews, document analysis, or observations, ideally in collaboration with a company, municipality, or platform provider.

Contribution:
The study will provide insights into how sharing economy practices can strengthen resilience in urban transportation and inform city governance and sustainable urban mobility strategies.

2025_024

This thesis aims to explore the potential of Artificial Intelligence (AI) to streamline and automate order processing workflows within a large-scale transport and logistics operation. Veøy is one of the largest privately owned transport companies in Norway, with over 450 trucks, more than 700 employees, and operations across Norway, the Nordics, and Europe. The company handles a significant volume of daily bookings and transport orders originating from multiple sources, including EDI messages, emails, and phone calls.

Currently, the process of receiving, classifying, and assigning these orders to the correct vehicle, route, and departure schedule relies heavily on manual routines. This can be time-consuming, prone to error, and resource-intensive. The objective of this research is to investigate whether AI-based solutions—such as Natural Language Processing (NLP), machine learning classification models, and automated decision-support systems—can improve efficiency, accuracy, and scalability in this critical part of Veøy’s operations.

The study will begin with a detailed mapping of the existing order-handling process, identifying bottlenecks, manual decision points, and opportunities for automation. Data availability and quality will be assessed, including EDI transaction logs, historical order data, and operational performance metrics. A conceptual AI-driven framework will then be designed, outlining how data can be ingested, processed, and used for automated assignment and scheduling decisions.

The research will also address key challenges such as system integration with existing transport management systems, data governance, and organizational readiness for digitalization. A feasibility analysis, potentially including a prototype or proof-of-concept model, will evaluate the expected impact on key performance indicators such as lead time, on-time delivery, and administrative workload.

The thesis ultimately aims to deliver actionable recommendations for how Veøy can implement AI-supported order processing to reduce manual work, increase operational reliability, and improve customer service.

2025_025

This thesis investigates how Artificial Intelligence (AI) can be applied to optimize workforce planning for a large-scale transport company. Veøy employs between 500 and 600 drivers whose work schedules must be planned weeks in advance while satisfying multiple and often conflicting constraints. These include compliance with EU and Norwegian regulations on driving and rest time, adherence to the Working Environment Act, alignment with collective agreements and company-specific agreements, and fulfilment of strict customer delivery requirements.

Currently, this planning process is largely manual, relying on tools such as Excel spreadsheets and requiring significant administrative effort. The research aims to explore how AI-driven optimization models and decision-support tools can create more efficient, legally compliant, and predictable work schedules for drivers, while reducing manual workload and supporting fair distribution of shifts.

The study will begin by mapping the current planning process in detail, identifying key data sources, including historical schedules, route demands, vehicle availability, and regulatory requirements. A conceptual optimization framework will then be developed, using methods such as mathematical programming, heuristic algorithms, or machine learning to generate feasible and robust driver schedules.

Attention will be given to ensuring that the proposed solution integrates with existing transport management and HR systems, handles real-world complexities such as last-minute changes and absences, and remains transparent and auditable for both management and employees. The feasibility of implementing such a solution will be evaluated through a combination of simulation, scenario analysis, and stakeholder feedback.

The expected outcome of the thesis is a set of practical recommendations and a prototype scheduling model that can help Veøy transition from manual planning to a digital, data-driven workforce scheduling process. This will support better compliance, improved resource utilization, and greater predictability for drivers and customers alike.

2025_026

This thesis explores how Artificial Intelligence (AI) and advanced optimization techniques can be applied to improve daily fleet planning and route scheduling for Veøy, one of Norway’s largest privately-owned transport companies. Veøy operates around 450 trucks distributed across 11 branches, serving customers nationwide and internationally. Every day, transport planners must allocate capacity across hundreds of vehicles to ensure all customer orders are executed in line with agreed delivery windows and contractual obligations.

Currently, this “fleet puzzle” is largely solved through manual work, with significant time spent matching orders to available trucks and routes. The objective of the research is to develop a data-driven and automated decision-support framework capable of maximizing vehicle utilization, improving cost efficiency, and ensuring operational compliance with driver schedules and delivery commitments.

The study will begin by mapping the current planning process and collecting data from the company’s order management system (load board), including daily demand, historical traffic patterns, vehicle availability, and cost structures. A core focus will be the design of an optimization engine that can:

• Generate optimal daily delivery plans based on factors such as real-time traffic data (roadworks, congestion,      accidents), fill rates, and economic performance of each route.

• Account for driver schedules and working time regulations, ensuring compliance while balancing capacity utilization.

• Provide predictive insight for future resource needs by leveraging historical data on seasonal demand patterns and vehicle downtime (e.g., sick leave or maintenance).

The research will also consider scenarios where traffic planners manually design routes but receive AI-powered feedback about potential inefficiencies, such as routes being too long, underperforming financially, or exceeding driver work limits. A feasibility study, possibly including simulation or prototype development, will be conducted to estimate potential improvements in utilization rates, profitability, and planning time.

The thesis aims to deliver actionable recommendations and a conceptual framework for implementing AI-supported fleet and route optimization, helping Veøy achieve near-100% utilization of its capacity while reducing manual workload and improving decision quality for planners.

2025_027

Background:
The Norwegian seafood industry is one of Norway’s most important export sectors, characterized by high requirements for quality, traceability, and sustainability. The transport from fishing vessels to landing facilities, through processing plants and subsequent transport stages, forms a complex supply chain influenced by multiple variables — including choice of transport mode, temperature control, time management, port infrastructure, coordination among actors, and regulatory as well as sustainability considerations. In addition, increasing documentation requirements related to climate footprint, efficiency, and food safety call for both digitalization and improved collaboration between actors in the supply chain.

Research Problem:
How can supply chains for transporting fish/biomass from vessels via landing facilities and transporters to the market be optimized in terms of efficiency, cost, quality, and sustainability?

Research Questions:

1. What are the main challenges in current logistics processes for transporting fish/biomass from catch to market?

2. How do different transport variables (mode of transport, frequency, lead time, route selection, temperature control, etc.) affect the efficiency and quality of the logistics chain?

3. How can digitalization and data sharing between vessels, landing facilities, and transporters improve coordination and reduce waste and climate impact?

4. What measures can be implemented to balance cost efficiency with the requirements for food safety, traceability, and sustainability?

Method:

• Case study of a selected value chain (e.g., pelagic fish or whitefish). The pelagic segment is the most accessible for data collection, given existing projects and regulatory requirements, while the whitefish segment—though somewhat more advanced—is also highly relevant.

• Data collection: Interviews with shipowners, landing facilities, transporters, and exporters; collection of transport and production data; document review.

• Analyses: Process mapping (value stream mapping), quantitative analysis of transport data (lead time, temperature, cost, CO₂ footprint), and scenario analyses of alternative logistics models.

Expected Contributions:

• Identification of bottlenecks and improvement opportunities in the transport segment from vessel to market.

• Recommendations for improved coordination and digitalization of transport logistics.

• Insights into how transport variables affect end results in terms of quality, efficiency, and environmental performance.

• Recommendations that can strengthen competitiveness and promote more sustainable seafood logistics.