This portfolio contains the code produced for the GEOG5995M module, which is saved in this repository. The repository evidences the different stages of the development of the final Agent Based Model.
The following table shows the correspondence of the files and the different practicals. Comments have been added to describe the code in blocks as appropiated. In some cases, lines used for testing purposes have also been commented to avoid the execution while running the code.
| Practical | Associated File(s) in the 📦Repository |
|---|---|
| 01 Agent Based Modelling | ┣📜01A_Agent_based_modelling.py ┣📜01B_Randomise_100by100.py |
| 02 Code Shrinking I | ┣📜03_ACode_Shrinking_II.py |
| 03 Code Shrinking II | ┣📜03_BCode_Shrinking_II.py |
| 04 Building Tools | ┣📜03_CTimingII.py |
| 05 Agents! 06 I/O |
┣ 📂model ┃ ┣📜agentframework.py ┃ ┗📜model_v0.py |
| 07 Communicating | ┣ 📂model ┃ ┣📜agentframework.py ┃ ┗📜model.py |
| 08 Animation/Behaviour | ┣ 📂model_Animation ┃ ┣📜agentframework.py ┃ ┗📜model.py |
| 09 GUI/WEB scrapping | ┗ 📂model_GUI ┣📜agentframework.py ┗📜model.py |
The ABM recreates the interaction of a specific number of individuals (agents) with each other and a defined environment. While all individuals move across a space limited by the size of the environment, their interaction with the environment consists of two actions, they can consume the environment's resources at a fixed rate; and, once their capacity is reached, deposit everything they have consumed. On the other hand, there is only one mechanism for interaction among individuals; they can share their resources with individuals located within a distance range.
The following sections describe the basic steps to open the Python files, set initial parameters, run the ABM, and obtain the outputs from the model.
The ABM was developed using the Anaconda distribution with Python 3.7. In order to run it correctly the following modules should be available in the Python environment:
matplotlib
tkinter
random
csv
requests
bs4
The files can be opened and run using a Python IDE such as Spyder or calling the model from the Anaconda command prompt. The second method requires less steps and, therefore, is less likely to fail. The steps are:
- Download the files of the model,
extract the model files from the
model_GUIfolder into a folder in your local drive. - In the Anaconda command prompt, change the working directory to the directory where the files are saved. For example:
cd /d M:\MSc\GOEG5995M\01_practical_agent_based\model_GUI
- Open the the files using the command
python model.py
A window will appear after a few seconds.
The GUI allows the user to the model parameters and run the model.
Three parameters can be specified using integers:
- the Number of agents in the simulation
- the distance threshold used to determine when the agents share their resources i.e. are in the same Neighbourhood
- the maximun Number of iterations to be produced
Once the parameters are defined, they can be set using the button on the top right. If one or more parameters are invalid or missing, default values are used.
To run the model, click on Model >> Run Model.
The window will show an animation with all iterations. In the plot, resources in each location are shown using a colour scale, and the size of the agents indicate the current amount of store.
Once the model has stopped, two outputs are produced in the same working directory.
FinalEnvironment.txtsaves the status of the environment after the selected number of iterations- Each line in
AgentStorage.txtcorresponds to the amount stored by all agents at the end of a single run of the model.