Here you can find the report of my first research internship.

The user manual for installing the necessary software and steps for running the models.

The System Dynamics project, created in AnyLogic 7.

The Cellular Automata source code, written in Java.

The Agent-based Modelling and Simulation source code, written and compiled in Java in Repast Simphony 2.0.

• After studying the Vensim PLE tool for mathematical modelling of the flu epidemic, I have found already well established model for epidemic dynamics. It is called the SIR model which is created in 1927 by Kermack and McKendrick. This model divides the population in 3 groups: S(susceptible) I(infected) R(recovered) and it is widely used for modelling diseases with usage of differential equations. The differential equation for the SIR model can be found here. I will try to implement the SIR model in Vensim PLE and model the flu epidemics, with the possibility of small changes or modifications of the SIR differential equations.

• The Bibliography page is updated with the most relevant papers I have read and needs to be updated periodically with the new papers.


• Written the summarizations for some of the most relevant papers in the fields of Agent-based Modelling and Simulation, Cellular Automata and System Dynamics.


• Started working with Vensim, but still collecting statistical data for the regular influenza in order for the models to be accurate and relevant. Important for comparing the results from the different models in different formalisms and with real world data. Also the statistical data is very important when working with Vensim PLE (System Dynamics), because without the real "numbers" the equations in the model will be inaccurate and the model will not be representing and simulating the real world as much as possible.


• The use cases for the different entities need to be changed after finishing with collecting the most accurate influenza statistics, for example: 1.Mortality rate? 2.How long does the influenza season lasts? 3.How fast the flu is spread? 4.When does the symptoms start? 5.For how long one person is infectious to another? etc. Some of these are already answered, but some need more research and more reliable resources.

• Found a Cellular Automata modeling and simulation tool called TerraME. This tool has an Eclipse plugin called TerraME-GIMS, which can be used to visually develop models. More research need to be done and the documentation should be read, before deciding to use this tool.


• The Use Case file for the model entities can be found here. Still some cases need to be added and some need to be refined, which will be done while simulating the agent-based model in the different tools. Link to the resources about the number of doctors per population in the EU.

• Nearly finished with the Java model of the flu epidemic in RepastSimphony 2.0. Still some refinement need to be done and also graphical representation of the results need to be created, which is possible in the runtime environment in RepastSimphony.


• Created a Use Case scenarios for the agents in the model, which explains what their behavior will be like and which actions they will execute. For now there are 3 different entities in the model: Healthy, Infected population and Doctors.


• Started with reading the documentation and testing the Vensim PLE tool, which will be used to model the same Use Case scenarios, but in the domain of Forrester System Dynamics. This tools uses mathematical equations and graphical representation of them, to help the user model and simulate different environments and scenarios. Still TODO: model and simulate the same Use Case scenarios created for the Agent Based Model and Simulation in RepastSimphony in Vensim PLE and compare the results.


• Created a brief skeleton structure of how the report should look like, but still the report should be started after the completion of the model and simulation in Repast. At the start there will be Abstract followed by Introduction, where the motivation and the problem will be defined. In the report there will be 3 different parts according to the levels of Abstraction, short summary about each of them, followed by the results obtained. There will chapter for comparing the results between the different levels of abstraction and also with real world data. At the end Conclusion and References.

• Reading and research in the 3 major fields (1. Forrester System Dynamics - Population Dynamics; 2. Cellular Automata; 3. Agent-Based Modelling and Simulation) contained as part of the Multi-Abstracttion Modelling and Simulation problem. Information about the relevant papers/articles/books can be found here(TODO bibliography).


• Researching and identifying some of the tools used for modelling and simulation. For now mostly ABMS (Agent-Based Modelling and Simulation) tools as Repast, AnyLogic. Other tools for modelling and simulation with Cellular Automata need to be reviewed and tested.


• As a starting point I have started building a small non-complex Agent-Based Model where infected people spread the flu when in contact with healthy population. For the construction of the model the Repast Simphony 2.0 tool is used which is a Java-based modelling system which also provides a simulation environment. Work is still in progress, the model needs to be refined and also the tool offers possibilities which need to be tested.


• I have found the tool Vensim which can be used to model and simulate System Dynamics in our case Population Dynamics during a flu epidemic. The tool offers free version Vensim PLE which can be used for personal and academic purposes. Still needs to be tested are in order to observe the results and compared with other techniques and tools.