On February 16th, 2024 the Norwegian Operations Research Society (NORS) held its Annual Master Thesis Award Ceremony at SINTEF in Oslo, celebrating together with more than 30 people in the audience the outstanding works of two exceptional scholars in the field of operations research: Marie Lindland and Helle Hagli Sønnervik. Gavin Bell, chairman of the NORS board, was happy to hand over flowers and certificates to the winners and their supervisors Carlo Mannino and Peter Schütz, respectively.

The NORS community extends its heartfelt congratulations for their exceptional achievements. Their theses have demonstrated excellence across multiple dimensions, including professional quality, originality, and potential impact on future research and practice. We look forward to seeing the continued impact of their work! Short abstracts are added below.

Marie Lindland (Supervisor Carlo Mannino): Mixed Integer Linear Programming Formulations For A Resource Constrained Project Scheduling Problem

The Resource Constrained Project Scheduling Problem (RCPSP) is an NP-hard job scheduling problem. This thesis finds and compares mixed integer linear programming formulations for an extension to the standard RCPSP with the weighted tardiness objective. We experiment computationally on test instances inspired by real data to find a formulation that gives good and stable solver performance. A new Big-M formulation for disjunctive constraints is introduced. This ”permutation formulation” has limited purpose and needs further study, but nevertheless outperforms a time-indexed formulation on some large instances. We deduce cutting planes from the set of disjunctive, resource, precedence and tardiness constraints and inspect how adding them to a formulation impacts solver performance. A family of strong cutting planes is deduced from the precedence constraints, and for these we describe a separation algorithm and a separation heuristic. Furthermore, we prove that these inequalities along with general constraints suffice to describe the convex hull of a specific integral set. A Block Decomposition Heuristic is designed to find feasible solutions to RCPSP instances where the precedence relations give rise to chains. Overall, we find that a classic full time-indexed binary formulation is suitable for solving RCPSP instances of various sizes and constraint characteristics. This result also holds when our RCPSP is formulated as a rescheduling problem.

Helle Hagli Sønnervik (Supervisor Peter Schütz): Strategic Fleet Renewal of Norwegian Fisheries with Environmental Considerations

Under the Paris Agreement Norway has committed to reducing ghg by at least 55% by 2030 compared to 1990 levels. Additionally, Norway has a legally binding goal to become a low-emission society by 2050, aiming for a 90-95% reduction in emissions compared to 1990 levels. Norwegian industries, including the fisheries sector, face increasing pressure to reduce their climate impact. As a significant contributor to Norway’s total CO2 emissions, Norwegian fisheries play a crucial role in achieving national climate goals. To meet these challenges, a comprehensive effort is needed to implement low- and zero-emission solutions in the Norwegian fishing fleet. ZeroKyst is a project that aims to contribute to a 50% emissions reduction from fishing and aquaculture vessels by 2030. This is accomplished through the development of a hybrid zero-emission powertrain and vessel that incorporates battery and fuel cell technology. For this master’s thesis, we formulate and solve a mathematical optimization problem for the strategic renewal of the Norwegian fishing fleet in order to provide decision support to decision-makers. We formulate the deterministic Fishing Fleet Renewal Problem with Emission Constraints (FFRPEC), which takes into account emission reduction targets for the period 2023-2050. The objective is to minimize the discounted total costs associated with fleet renewal and operation of the Norwegian fishing fleet. From solving the model, we obtain a detailed schedule specifying the timing of replacing a certain number of vessels with a specific propulsion system within a sub-fleet, as well as the propulsion system to be used as a replacement. The parameter values used in the model are calculated using data regarding the existing fishing fleet and cost information associated with propulsion system components and associated energy storage, such as combustion engines, batteries and fuel cells. We conduct an extensive scenario analysis on uncertain parameters such as the fuel prices of mgo, battery power, hydrogen and ammonia, CO2 tax price-trajectories, costs of the mentioned propulsion system components, as well as shipyard capacities. The analysis aims to identify the potential impact of the relevant parameters on the resulting fleet renewal schedule. We highlight three key findings from our analysis: (1) Zero-emission propulsion is economically unfavourable (2) Penalizing the use of conventional fuel incentivizes earlier renewal of the fleet (3) Immediate action must be taken to initiate the renewal of the ocean-going fishing fleet to achieve emission reduction targets in a cost-effective manner.

UiB seminar

Speaker:

Dr. Matthias Winkenbach is the Director of the MIT Megacity Logistics Lab and a Research Scientist at the Massachusetts Institute of Technology (MIT) Center for Transportation & Logistics (CTL). He is also spearheading a new research initiative at MIT CTL at the intersection of supply chain and logistics, data science and visualization, and human decision-making – the MIT Computational and Visual Education (CAVE) Lab.

Title:

Modeling real-world urban logistics systems – using data to master the last mile

Abstract:

Urban mobility systems are facing a number of major challenges that threaten their future ability to sustain the economic and social activity of cities. First, urbanization is progressing at a rapid pace, with more than 85% of the global population expected to live in cities by 2030. Unprecedented levels of urban density, both in terms of population and economic activity, call for innovative mobility solutions for people and goods. Second, the boom in ecommerce and `on-demand consumerism´ impose an additional burden on urban logistics systems and the underlying infrastructure. Urban freight volumes are growing and becoming increasingly fragmented as customer expectations towards the speed, flexibility, reliability, and customization of their shipments are rising quickly. Existing planning tools and distribution approaches no longer allow for an effective consolidation of urban freight flows on cost-efficient and well-utilized vehicle routes. Third, the increasingly severe effects of urban mobility on the environment and public health require the adoption of cleaner, smarter and more efficient vehicles.
In this talk, Dr. Winkenbach will touch upon some of the quantitative methods and data sources employed by his team of researchers at the MIT Megacity Logistics Lab to solve real-world last-mile logistics problems, and to enable an optimal strategic design and operational planning of freight distribution systems in complex and volatile urban environments. Further, he will highlight the growing importance of data science methods for the optimal design and control of urban freight mobility systems. Using the example of several prototype applications under development at MIT CTL’s newly created MIT Computational and Visual Education (CAVE) Lab, he will demonstrate the potential of interactive information visualization for data- and analytics-driven supply chain and logistics decision making.