When: Monday, 06 May 2024, 09:00-17:00.
Where: Industriens Hus, H. C. Andersens Blvd. 18, 1553 København, Denmark.
Applications of Optimization 2024 (AOO 2024), our annual conference on applied mathematical optimization takes place on Monday, 06 May 2024 at Industriens Hus in Copenhagen (09:00-17:00).
Sponsors: MOSEK, ØRSTED
Below you find the programme, followed by the abstracts of the conference talks.
Programme
The workshop program includes six talks from OR experts and practitioners, and there will be ample time for networking:
09:00 – 10:00 Registration and refreshments
10:00 – 10:10 Welcome and announcements
10:10 – 10:50
Fredrik Odegaard, Associate Professor, Management Science, Ivey Business School
Title: Intra-Provincial Benchmark Analysis of COVID-19 in Canada; Efficiency Measurement with Desirable and Undesirable Factors
10:55 – 11:35
Kevin Tierney, Professor for Decision and Operation Technologies, Bielefeld University
Title: Deep reinforcement learning techniques for solving combinatorial optimization problems
11:35 – 11:45 Sponsor pitches
11:50 – 12:50 Lunch
12:50 – 13:30
Filippo Focacci, PhD, Co-founder and CEO, Decision Brain
Ella Pirttikangas, Business Analyst, Decision Brain
Title: Maintenance and Workforce Planning and Scheduling with ML & Mathematical Optimization
13:30 – 14:10
Rikke Kristiansen, Senior Simulation Modelling Engineer, Shoreline:
Title: Optimizing maintenance of offshore wind farms
14:10 – 14:40 Coffee break
14:45 – 15:25
Mathias Hintze, Analyst, Mærsk Mc-Kinney Møller Centre for Zero Carbon Shipping and
Frederik Lehn, Model Developer, Mærsk Mc-Kinney Møller Centre for Zero Carbon Shipping:
Title: Maritime Decarbonization - Simulating the maritime industry’s transition to net-zero emissions
15:30 – 16:10
Rafael Barfknecht, PhD in theoretical Physics and Quantum Engineer, Kvantify
Title: Prospects of quantum computing in optimization and related tasks
16:10 – 16:20 Closing
Speakers and Abstracts
Fredrik Odegaard, Associate Professor, Management Science, Ivey Business School
Title: Intra-Provincial Benchmark Analysis of COVID-19 in Canada; Efficiency Measurement with Desirable and Undesirable Factors
Abstract:
In this talk I will present the findings of an optimization-based efficiency analysis of the Covid-19 pandemic in Canada. Like all countries, the COVID-19 pandemic posed unheralded challenges to the people, business, government at all levels (federal, provincial, regional), and society at large in Canada. In addition to the direct consequences of taking care of infected people, the pandemic strained eldercare, employment, economic growth, and exacerbated mental health and social problems. During the first year of the pandemic, generally speaking, researchers' and policy makers' main focus was on “flattening the curve,” and on predictive modeling of infections and deaths. In this talk I will discuss a non-parametric data-driven analysis based on Data Envelopment Analysis to assess COVID-19 in the ten Canadian provinces over the two year period 2020 to 2021. The objective was to derive worst- and best-case intra-provincial benchmarks to assess if and to what extent the situation could have been worse respectively better. To take account for any indirect socio-economic impact the analysis incorporates official monthly unemployment rates and a stringency index reflecting the level of social policy restrictions imposed by the provincial governments. A major contribution of the model framework is that it provides a mechanism for measuring the impact of the two main strategies in curbing the pandemic, namely vaccination and social policy restrictions. As a robustness check, the bench-mark results are compared against bias-corrected efficiency measures. The methodology presented is not limited to the analysis of pandemics, but is completely general and can be extended to the efficiency analysis of any process that incorporates both desirable and undesirable factors.
Kevin Tierney, Professor for Decision and Operation Technologies, Bielefeld University
Title: Deep reinforcement learning techniques for solving combinatorial optimization problems
Abstract:
In recent years, deep learning techniques have made incredible progress in finding high-quality solutions to combinatorial optimization problems. But are they ready for real-world applications? I provide an overview of the latest deep learning-based optimization methods and contrast the progress in this field with current state-of-the-art Operations Research (OR) heuristic solvers. This talk thus offers a critical assessment on whether these methods beat "traditional" OR approaches or not (spoiler: they do not -- mostly) and gives an outlook on where the field of "neural combinatorial optimization" is headed.
Filippo Focacci, PhD, Co-founder and CEO, Decison Brain
Ella Pirttikangas, Business Analyst, Decision Brain
Title: Maintenance and Workforce Planning and Scheduling with ML & Mathematical Optimization
Abstract:
In this talk I will provide an overview of different use-cases and methods used for planning and scheduling in the field of Maintenance and Workforce optimization highlighting challenges and possible approaches.
Rikke Kristiansen, Senior Simulation Modeling Engineer, Shoreline
Title: Optimizing maintenance of offshore wind farms
Abstract:
Shorelines Design tools make it possible to simulate the construction and maintenance of offshore wind farms from anywhere between a couple of months to over 30 years.
During this time, we need to schedule the work on the different turbines daily, involving multiple vessels to transport personnel and many different turbine locations, each with tasks with different severities.
In this talk, I will describe how this has been solved, considering the demand for fast results and some of the more complex requirements coming from working on offshore turbines, such as strict weather restrictions and safety rules.
Mathias Hintze, Analyst, Mærsk Mc-Kinney Møller Center for Zero Carbon Shipping and
Frederik Lehn, Model Developer, Mærsk Mc-Kinney Møller Center for Zero Carbon Shipping:
Title: Maritime Decarbonization - Simulating the maritime industry’s transition to net-zero emissions
Abstract:
Join us to learn how we are supporting maritime decarbonization by modeling likely transition scenarios towards net-zero emissions for the industry. Shipping is a hard-to-abate sector and consequently requires new types of fuel and technology to decarbonize. These fuels and technologies have different cost and availability outlooks. Further, the sector is impacted by policy decisions and market dynamics. To combine insights across all the value chain, the Mærsk Mc-Kinney Møller Center for Zero Carbon Shipping has developed a simulation model, NavigaTE.
NavigaTE is designed to simulate the transition of the maritime industry. This is done by modeling the decision processes of the different actors along the maritime value chain. A core part of NavigaTE is the fuel selection algorithm which is solved at every time-step using a Linear Programming (LP) Solver.
We will discuss a few topics including:
What decarbonization in shipping could look like.
How the maritime industry can leverage modelling and data-driven decisions in its transition to net-zero emissions.
How NavigaTE is designed and used to simulate decisions across a diverse group of stakeholders.
How an LP solver is used in an unconventional way as part of a larger simulation algorithm.
Rafael Barfknecht, PhD in theoretical Physics and Quantum Engineer, Kvantify
Title: Prospects of quantum computing in optimization and related tasks
Abstract:
Quantum computing represents a new way to process information using concepts such as quantum superposition and entanglement. This opens up for new types of algorithms, both those that are run on purely quantum hardware, and those that leverage a mix of classical and quantum hardware.
In this talk, I will start with an introduction to the central concepts of quantum computing, including some key quantum algorithms. I will then proceed to discuss use cases in optimization and some outlooks on what is ahead as the hardware matures.
Comments