Original README.md can be found here with some informations about how the system works.
This agent was developed base on the following repository provided as a tamplate for 2021/2022 edition of 01URROV - Computational Intelligence at Politecnico di Torino.
For launch the server the basic syntax is:
python3 server.pyIs an hard coded implementation, with a focus on the discard move, the aim of the agent is to avoid wrong moves (for a wrong move is meant to use a red token).
For launch the client the complete syntax is:
python3 client.py <SERVER_IP> <SERVER_PORT> <CLIENT_NAME>If you launch python3 client.py: the default values for the parameters are:
SERVER_IP = 127.0.0.1, SERVER_PORT = 1024 and CLIENT_NAME = SuperPippo
After the launch of the client, you have to put the following command: ready for start the communication with the server and to tell to the others players that the client is ready. If you want to close after ready command use the combination of CTRL+C because the input is not read anymore (in the template implamentation the input was managed asynchronous), before giving ready command also exit command is available for quit the client.
For developing the hard coded version the only support used is a translated version of the rules, available here, and numpy online communities.
All the "intelligence" of the agent is inside the suggestMove() function, where after asking for the current state of the game to the server is computing which play is more suggested.
The agent's hand is matrix based on the hints from others players or from moves done by the agent it self.
This works was developed by Mattia Dutto (s287598@studenti.polito.it)