Agent Based Computational Economics SE ST 2020

The course is stuctured as a literature seminar, which deals with selected topics in a quite modern branch of Computational Economics - Agent Based Computational Economics / Modelling (ACE).
This type of modelling employs bottom-up techniques that focus on the interaction between heterogenous indiviual agents (households, firms,etc.) and their environment - based on behavioral rules. This way, various important nonlinear processes can be studied and phenomena investigated and explained, which are heavily influenced by the heterogeneity of the individuals (non-uniform distributions of attributes - age, income, wealth, etc.), their evolving specific interaction structures (social networks) and their often boundedly rational decision making, which could otherwise not be explained and understood.

The work assignment for students consists of multiple parts: (0) coming to the initial meeting, (1) preparation of a presentation (including slides) for a talk presenting the paper, (2) giving the actual talk and answering to following questions, (3) attending the other presentations, listening and actively participating in discussions following the other presentations and finally (4) preparing a written report (essay) describing the paper using your own words.

The available presentation dates will be/have already been discussed/determined during the inital meeting. Your personal presentation date is determined by your choice/registration of available dates in the TISS system.

Please register for the course in TISS and attend the initial meeting!

After you have chosen 2-3 topics (scientific papers) from the list of available topics, please send a e-mail with those 2-3 preferred topics (in preference order). Once I reply by stating which of these topics was still free and which one is now assigned to you, you may immediately start working on the written report and preparing the presentation.

The written report has to reach me (via E-Mail) by no later than 1 month after the end of the semester. It should be approx. 13-15 pages long (net length - i.e. without cover/back sheet, table of contents, table of figures, references, acknowledgements, etc.). The number of pages is understood with either Arial,Verdana, Calibri [and comparable san-serif fonts] size 11, 1.5 line-spacing, or TimesNewRoman [and comparable serif fonts] size 12, 1.5 line-spacing.

You will have approx. 25-35 minutes time for your talk - please plan your presentation accordingly and send me your slides by E-Mail at least 3 days prior to your talk. Please not that it is of course also necessary to be present during all other presentations (on all dates) and to participate in discussions following presentations.

The focus of your efforts should be to produce a comprehensible rendition of the paper by thoroughly and critically reading the paper and thinking about the discussed topic. It is important to include all relevant points (not every single equation if the model is described in mathematical form exhaustively) and yet present the paper - in the available speaking time - to your colleagues who have not read the paper in such a way that they can still follow and understand your presentation. Observing the time limit for your presentation is not only a question of courtesy, but also a kind of quality criterion of your work, because correct and efficient summarizing requires a much higher degree of comprehension than simple repetition does. Furthermore, it is key for professional presentations - especially scientific presentations (e.g. on conferences), where talking time is often severley limited (something between 10 and 30 minutes) and sometimes even further reduced on spot.

As the course's language of instruction is English, both the written report and the presentation have to be held in English. Though mastery of the English language is not a grading criterion, incomprehensible formulations or probably wrong explanations will of course have a negative influence on your grade.



Available Topics

If you choose a very short paper (below 15 pages) you have to research additional information (an additional paper or parts from other scientifically acceptable sources) and include it in your report and presentation.

The following papers are chooseable topics in 2020:


Macro-economic ABM (ABM of the Economy) topics:

To bail-out or to bail-in? Answers from an agent-based model
topic assigned to student in ST2020

Banks, Market Organization, and Macroeconomic Performance: An Agent-Based Computational Analysis

Unemployment benefits and financial leverage in an agent based macroeconomic model

Evolutionary macroeconomic assessment of employment and innovation impacts of climate policy packages
topic assigned to student in ST2020

Agent based macroeconomics - a baseline model
topic assigned to student in ST2020

Various socio-economic ABM topics 2020:

An agent-based model of school closing in under-vaccinated communities during measles outbreaks

Modeling learning and forgetting processes with the corresponding impacts on human behaviors in infectious disease epidemics

Epidemic Spreading in Urban Areas Using Agent-Based Transportation Models
topic assigned to student in ST2020

Simulating the spatial diffusion of memes on social media networks

Coordination in Transient Social Networks: An Agent-Based Computational Model of the Timing of Retirement
topic assigned to student in ST2020

Tragedy of the commons



Hands-on papers for ACE/ABM

The following papers are of interest if you want to replicate or develop a simulation yourself (pure interest, project course, diploma thesis, etc.) but they can't be chosen as topics:

How to replicate an Agent-Based Model

How to discern results from different: Errors and Artefacts

Visualization of Simulation Runs and Results
(Analyzing and presenting simulation results is a very important point (multi dimensional results))

An extensive overview of different simulation frameworks/programs
(Though for most questions Netlogo, Repast 3.1 and Repast Simphony are the most appropriate systems, most other programs)

For some research questions empirical validation of results is crucial:
Empirical Validation of Agent Based Models

Alternative approaches to Empirical Validation

Micro- and Macro-Level Validation in Agent-Based Simulation:
Reproduction of Human-Like Behaviors and Thinking in a Sequential Bargaining Game




Introductory Papers

The following papers are only suitable as introductions to computational economics / agent based simulation / complexity in economics / bounded rationality, they also can't be chosen as topics.

Agent-based Computational Models and Generative Social Science

Remarks on the foundations of agent based generative social science

Computationally intensive analyses in economics

Agent Based Computational Economics - a research survey

The economy as an evolving complex system

Out of equilibrium economics and agent based modeling

Inductive reasoning and bounded rationality - The El Farol Problem

 

Interesting Links

One of the most interesting resources is the online-only journal JASSS (all papers can be read/downloaded free of charge):
Journal of Artificial Societies and Social Simulation

Prof. Leigh Tesfatsion's ACE Website with maaaany ACE information, material and further links:
Leigh Tesfatsion's ACE Website