Autonomous Software Agents Project

This project is part of the "Autonomous Software Agents" course of "Artificial Intelligence Systems" master degree.

The objective is to develop an autonomous agent based on the Belief-Desire-Intention (BDI) architecture. The agent is designed to play Deliveroo.js game.

Project Structure

├── Beliefs
│   ├── map.js              # Manage informations about the map
│   ├── beliefs.js          # Manage agent's beliefs
│   └── sensing.js          # Manage agent's perceptions of parcels and other agents 
├── Communication           # Multi-agent mode
│   ├── communication.js    # Manage communication between agents 
│   └── teamOptions.js      # Manage the communication of options between agents
├── deliveroo               # Game configuration
│   ├── client.js
│   └── config.js
├── index.js
├── Intentions
│   ├── intentions.js   # Intention Execution
│   └── options.js      # Compute utility and choose best option
├── main.js
├── Plan
│   ├── astar_search.js # Search algorithm
│   ├── PDDL            # Contain PDDL domain and problem files
│   └── plan.js         # Manage all possible plans
└── utils.js

Intallation

Clone the repository:

git clone https://github.com/MartinaPanini/BDI_Agent.git

Install dependencies: npm install Setup environment In index.js:

const host = "https://deliveroojs2.rtibdi.disi.unitn.it/";
const mexican = { id: '', name: '', token: '' };
const caravan = { id: '', name: '', token: '' };

Set the agent's parameters: id, name and token. In alternative, you can copy the token in index.js and past in Deliveroo game during the login phase.

Running

• Connect the agent(s) in Deliveroo game
• Go inside BDI folder
• Single mode: node index.js single
• Multi-mode: node index.js multi