A semester-long group project split into 3 phases. Please consult 2021-2022_2-2_ProjectManual.pdf for more information about the project and its phases.
Our code for each phase is in a separate folder and inside there's a README file that explains how to run the code.
Short summary: We created a multi-agent surveillance game and implemented A* path-search algorithm and Q-learning into it. For research, we examined how the effects of the variations of critical parameters and agent sets impact general learning performance. The research questions aim to determine the effects of the implemented algorithms on maps of different complexities.
For further explanation of what we did, please consult FinalReport.pdf.
Group members: Alisa Todorova, Cem Ayder, Tiphanie Bent, Collin Makuza, Zofia Milczarek, Zaker Omargeel, Vikram Venkat