Simulation Environment for Electric Taxi Management

This environment uses Tkinter to render the GUI and it is meant for testing different reinforcement learning algorithms.

Installation

First, make sure that you have the right tkinter version that goes with the python version you are going to use. And then install the requirements with pip:

pip install -r requirements.txt

Implemented Algorithms

• DQN (closely follows the DQN implementation from this amazing repository)

Environments

• Grid Environment:
  • environment:
    • Passengers = 1
    • Charging Stations = 1
    • E-taxi: 1
  • multi_env:
    • Passengers = 5
    • Charging Stations = 4
    • E-taxi: 1
• Graph Environments:
  • dispatch_env:
    • Variable Number of passengers
    • Charging Stations = 2
    • E-taxi: 1
    • Better control over parameters of EV
    • Real Network topology, traffic conditions and charging station locations

Example

Run min_run.py to see a preview of the Graph Environment for a single episode. This script has most of the information necessary to run the environment.