Project made for Autonomous Agents and Multi-Agents Systems course. It consists on a simulation of a mask distribution service, with the objective of better understanding the impact that masks can have in the virus’s spread and to analyse the best strategies that can be used for performing such a task. It has three different implementations for the agents responsible for distributing the masks: reactive without internal state, reactive with internal state and reinforcement learning based.
- Open the MaskDelivery.exe file;
- In the menu, tick one of the boxes according to the agent you want (not ticking will result in a reactive agent with no internal state):
- RL agent: agent trained using Reinforcement Learning;
- Internal state: purely reactive agent with an internal state;
- Insert the number of nurse, infected and patient agents;
- Choose the probability of an agent getting infected by using the slider;
- Hit "Play" and click on the "Hide" button to watch the simulation run.
Link -> Video