Below you can find the preliminary program of the conference and the workshop’s preliminary description. More details about the keynote talk, the workshop descriptions, their agendas, and exact rooms will be provided soon.

Even though the talks are already planned, we encourage everyone to present their current research during the poster session.

Agenda

Workshops

Fantastic Models and How to Make Them: A Beginner’s Guide to CaterFly

Meet CaterFly, a community that brings together early-stage researchers who work at the intersection of agent-based modelling and behavioural science, with the energy transition as their main topic of interest. Next to battling the difficulties of conducting interdisciplinary science, we aim to learn with and from each other, sharpen our research together, and enhance collaboration in this new and upcoming field. CaterFly was established in the spring of 2023, it spawned from a small network event. With members from well over 10 countries, it is growing ever since with monthly talks and small-group projects.

Description of workshop exercises

Activity one – Rock Paper Scissors (30 min)

This part aims to show the impact of slight variations in behavioural rules of agents. We first play a game of rock-paper-scissors in duos and break down the simple rule set of agents together. We then play a group game of rock-paper-scissors with two different roles: hunters and prey. After this group game, we brainstorm on the personal rules we employed to win the game. We compare this to a formalised implementation of a pre-made ABM. This will demonstrate how different implementations of behavioural rules and insights significantly affect a model’s outcome – and how modelling choices eventually depend on models’ purpose and modellers’ intent.

Activity two – Mapping our work (45 min)

Having gained a first insight into the challenges regarding the interface of ABM and behavioral science, we turn our attention to the people who deal with these challenges. To this end, a visual overview is co-created by participants to illustrates the current approaches in participants’ own research on integrating behavioural science in agent-based modelling. Participants first position themselves randomly in the room. Then, they relocate based on 1-minute research pitches and the discipline they are trained in. With each relocation, participants move closer to the speaker or further away based on the level of interconnection with their own research. The final stance of participants in the room is recorded on a digital map, including key information on their research.

Target audience

Early-stage researchers working at the intersection of agent-based modelling and behavioural science, meaning the incorporation of some type of human behaviour into agent-based modelling.

Preparation for participant

Freshen up your rock-paper scissors skills
Prepare a 1-minute pitch regarding your research that combines ABM and behavioural science

Presenters

Mariëlle Rietkerk is a PhD candidate at the Energy Transition Lab of TU Delft’s faculty for Technology, Policy, and Management. She focuses on the implementation of behavioural insights and empirical data in ABMs. She has a specific interest in formalizing different ‘hassle factors’ to energy behaviours of households (e.g. adoption behaviours like buying solar panels or heat pumps and habitual behaviours like load shifting). Mariëlle holds a master’s degree in Economic and Consumer Psychology from Leiden University (2015). She worked as a social scientist at TNO and Milieu Centraal before she joined TU Delft.

Lukas Schubotz is a PhD candidate at the Energy Transition Lab of TU Delft’s faculty for Technology, Policy, and Management. In my work, I focus on the computational aspects of scrutinising ABMs for the energy transition, using inverse modelling and machine learning to explore behavioural theory or policy effects through modelling. I hold a bachelor in mathematics with a minor in psychology from Heidelberg University, and a master in innovation science from Utrecht University. Up for all types of academic banter and creative ideas!

Hannah Galster is a PhD candidate at the Copernicus Institute of Sustainable Development (Utrecht University) and the Energy and Materials Transition unit of TNO (Netherlands Organization for Applied Scientific Research). Working on the incorporation of human behavior in energy system modelling, she aims at combining insights from social sciences with energy modelling. Hannah holds a bachelor’s degree in Psychology (Erasmus University Rotterdam) and a master’s degree in Environmental Sciences (Utrecht University).

Philipp Eppe is a research associate at the Department of Resilient Energy Systems at the University of Bremen and a PhD student at the Department of Environmental Psychology at the University of Groningen. Philipp studies the factors influencing the adoption of sustainable technologies, focusing specifically on corporate decision-makers. His research’s central goal is to formalize these findings into agent-based models to increase their psychological realism and support local transitions. Philipp holds a BSc in Psychology and an MSc in Environmental Psychology from the University of Groningen.

Hands-on Workshop: Using AMIRIS

AMIRIS is an agent-based electricity market model developed since 2008 and open-sourced in 2021. It provides an innovative approach to analysing and evaluating energy policy instruments and mechanisms for the market integration of renewable energies and flexibility options. Simulations with AMIRIS allow a broad range of analyses. For example, the influence of regulatory framework conditions on the behaviour and profitability of energy market players can be investigated. In addition, users are able to assess the impact of, e.g., future fuel or CO₂ prices on electricity price and technology-specific market values. Due to its agent-based and modular nature, AMIRIS can be easily expanded or modified. We aim to empower other scientists to employ AMIRIS, to participate in its development, and to adapt and extend it to their needs. Last but not least, we strive to create transparency for the modelling process and its results.

In this workshop, we will introduce the basic terms used in AMIRIS and shape a basic understanding of how agents interact. We will proceed with a hands-on installation of AMIRIS. After a first run of AMIRIS with an existing parameterisation of Germany 2019, we will take a look at the output and plot important results. We will then experiment with different input parameters (e.g. CO₂ prices, or remuneration policies) and assess their impact on the model outcomes. Finally, we will introduce the available documentation for the model and its code.

Requirements

Git
Java JDK 11 or higher
Python 3.9
See also the workshop-specific installation instructions.

Schedule

20 min – Basics on AMIRIS’ Agents
15 min – Installation and Execution
25 min – Result Analysis
30 min – Parameter Variations
10 min – Documentation

Presenters

Christoph Schimeczek completed his doctorate in physics in 2014. Since then, he has been working on agent-based modelling at the German Aerospace Center (DLR) – initially in the field of electromobility – and since 2017 in the Energy System Analysis department. Since 2019, he is responsible for the development of the agent-based modelling framework FAME and the electricity market simulation model AMIRIS. His research focusses on the modelling of flexibility options in energy markets the coupling of different models and model types in the energy systems domain.

Priv.-Doz. Dr. Dr. Ulrich Frey completed his doctorate in Philosophy of Science in 2007 and in Biology in 2016. In 2017, he habilitated at the University of Giessen, where he was awarded the venia legendi in environment and natural resources. He has worked as a research assistant at the University of Braunschweig, Giessen and Halle from 2005 to 2016, where he represented, among others, the chair of Prof. Voland in 2012/2013 and was a guest researcher at the Workshop for Political Theory and Policy Analysis in 2010. Theory and Policy Analysis. There, he worked with Nobel Prize winner Prof. Dr. Elinor Ostrom. His research topics and expertise include renewable energy, natural resource management and cooperation research, agent-based modelling, surveys, machine learning and various statistical methods.

BedDeM: Behavior-driven Demand Modeling

BedDeM is an ABM model that focus private households demand. The purpose is to investigate demands for energy-related services, as not only is technological acceptance required, but their adoption to make an impact. This is true, both for daily decision-making (e.g. mobility mode choice) but also demand for long-term goods (vehicles purchases).

In BedDem agents decision-making does not follow a particular economic theory, i.e. minimizing some cost function, which does not comprise for the complexity of the underlying decisionmaking process. Neither do they purely rely on statistical data, i.e. as in discrete choice models, which is not sound when several variables are changing at the same time. Instead, BedDeM agents rely on a psychological model for individual decision-making, which seems an appropriate start for modeling consumers behaviors. In particular we have selected Triandis theory of interpersonal behavior as our theoretical foundation. This architecture encapsulate that a variety of factors is typical for the decision-making, and that for each and every of these factors a potentially diﬀerent outcome would be optimal. Decision-making is happening by combining the diﬀerent factors. In addition we have integrated additional aspects to better capture bounded-rational eﬀects, which rely on a more detailed modeling of the agent, its’ perception and interaction with it’s environment.

In currently ongoing research we aim to use the model to further advance the grounding of ABMs, not only in terms of a strict bottom-up construction of the ABM, its data-driven calibration and validation, but also in measuring and modeling the precise impacts of interventions that are needs to be evaluated by the ABM, e.g. which people will react how to an intervention. This research is performed in a joint research eﬀort with psychologist.

From Challenges to Solutions: ASSUME’s Role in ABM for Energy Markets

Overview: During this workshop our focus will be on exploring the challenges and innovations in agent-based simulations through collaborative discussions, tailored lightning talks, and in-depth tutorials building on the discussion results. Participants will engage with the latest advancements of ASSUME, an open-source toolbox designed for simulating electricity markets using DRL algorithms.

Workshop Objective: This session aims to foster a collaborative environment for sharing knowledge, addressing the complexities of modelling modern energy markets, and exploring innovative solutions through the lens of ABM and DRL. By the end of this workshop, participants will have a deeper understanding of the challenges in agent-based energy market simulations, and practical insights into using the ASSUME framework.

Agenda

Part 1: Collaborative Discussion on ABM Challenges (50 minutes)

Introduction (5 Minutes): Who are we, you and what is this workshop about?
Group Activity (15 minutes): Participants will be divided into groups to identify and list 3-5 key challenges.
Group Discussion (15 minutes): Each group discusses their identified challenges, sharing insights and experiences.
Plenary Discussion (15 minutes): Discussion of the highlighted challenges and exploration of potential tools and strategies for addressing them.

Part 2: Lightning Talks and Discussions (40 minutes)

Introduction (15 minutes): Brief overview of ASSUME Framework
A series of 5–7-minute talks covering the most frequently mentioned challenges, followed by discussions on their possible solutions using state-of-the-art ABM modelling such as the ASSUME Framework.

Part 3: Tutorials (90 minutes)

This part of the workshop will provide participants with several tutorials on using the ASSUME framework. In particular, these tutorials will focus on the challenges in ABM addressed in the previous part and possible way on how they can be addressed with ASSUME. The topics include but are not limited to:

Deep Dive into ASSUME and Reinforcement Learning
Multi-market Simulation with ASSUME
Advanced Agent Configuration and Demand Side Management in ASSUME

Presenters

Nick Harder is a research associate and a doctoral student at the Institute for Sustainable Systems Engineering (INATECH) at the University of Freiburg. He completed his master’s studies in “Sustainable Systems Engineering” at the same university. His research primarily focuses on utilizing deep reinforcement learning methods to model electricity markets and understand the behavior of market participants. Additionally, he holds the role of project coordinator for the ASSUME project, which focuses on developing a modeling toolbox for analyzing electricity markets and market designs through the use of deep reinforcement learning techniques. Overall, Nick’s expertise lies in the intersection of sustainable systems engineering and advanced machine learning approaches applied to energy markets.

Kim K. Miskiw is a research associate and a doctoral student at the Chair for Information and Market Engineering (IISM) within the Faculty of Economics and Business Engineering at the Karlsruhe Institute of Technology (KIT). Her research interests revolve around deep reinforcement learning in electricity market simulations, agent-based electricity market modeling, energy market engineering, and stochastic optimization. Previously, she held the position of Junior-Project Associate at the Institute for Industrial Production, Chair of Energy Economics (KIT). Kim completed her Master’s degree in Industrial Engineering at KIT, focusing her master’s thesis on stochastically optimized bidding strategies in sequential electricity markets and examining their benefits in relation to risk preferences and portfolio setups.

Florian Maurer is a research associate and doctoral student at the University of Applied Sciences Aachen in cooperation with the University of Oldenburg. After completing a dual study program in software development, he obtained his Master’s degree in “Applied Mathematics and Computer Science” at FH Aachen, where he developed charging solutions for e-mobility. Florian is involved in research projects related to energy measurements and prosumer market integration. His research interests include open-source development, wireless communication and energy market design. Currently, he is researching agent-based modeling of energy markets to provide a simulation framework that covers the comparison of different market designs and policies.

Bidding strategies on electricity markets taking into account the expansion of renewable energies and European market coupling

In the BMWK-funded project “Analysis of bidding strategies and their influence on the electricity price in the EU internal market – BETS”, the project partners from the Karlsruhe Institute of Technology (KIT), Danmarks Tekniske Universitet (DTU) and the Institut für Zukunfts Energie- und Stoffstromsysteme (IZES) are analyzing strategic bidding on electricity markets. In the workshop, the initial research results on bidding strategies developed in the project will serve as an impetus for an in-depth exchange on this important topic. A system developed to categorize the strategies will be presented, discussed and, if necessary, supplemented. Finally, the strategies developed will be discussed with regard to mapping them in an agent-based simulation. Short keynote speeches will introduce the topic and open up group discussions.

Target audience: All those interested in the topic are welcome, both researchers (PhD students, PostDocs) on energy markets and on agent-based modeling.

Agenda Slot 1:

Warm-up in the form of a short game on the topic of bidding on the electricity market
Exchange / discussion / group work
Presentation of relevant regulatory framework conditions in Germany and selected European countries
Exchange / discussion / group work
Description of selected bidding strategies and corresponding categories I
Exchange / discussion / group work

Agenda Slot 2:

Description of selected bidding strategies and corresponding categories I
Exchange / discussion / group work
Wrap-up
Impulse presentation to decision making in agent based models
Exchange / discussion / group work
Wrap-up

Location

Kollegiengebäude I, Room HS 1010
University of Freiburg
Platz der Universität 3
79098 Freiburg, Germany

For the locations of lunch, dinner, starting point for the city tour, and others, check arrival information.