40.305 Project - Elevator Group Control

Project for 40.305 Adv Topics in Stochastic Modelling Based on Elevator Group Control Problem

For more information about the project report details, please go to the wiki

Installation and running the simulation

Requires Python >= 3.10 Create a new virtual environment and activate it:

python -m venv venv
./venv/Scripts/activate

Install the requirements:

pip install -r requirements.txt

Running the simulation

To run the simulation, simply run

python main.py

from the main project folder

Various argparsers will be implemented soon to specify features of the simulation

Code for the analysis can be found in /analysis.

Running interactively with Jupyter Notebook

A sample jupyter notebook is shown in interactive_analysis.ipynb The simulation can be imported as shown in the notebook.

Objectives to Minimize

• Wait Times of passengers by floor
• Idle Times of each elevator

Variables Tested

Controller Type

• Random Assignment (Control)
• Zoning (Sectoring)
• Nearest Elevator (Feasibility Score)

Idle Floor Configuration

Various Arrival distributions

To test the limits of various Idle Configurations, we experimented with different kinds of arrival distributions Namely

• Uniform Arrival Distribution
• Ground Floor Heavy Arrival Distribution

Uniform Arrival Assumptions

• Assumes that an equal amount of people arrive at every floor
• Equal proportion of people want to go from one floor to every other floor

Ground Floor Heavy

• A fixed proportion of people arriving and leaving every floor
• An equal proportion of people go from ground floor to every other floor
• Same proportion of people going from ground to other floors is same as proportion going from other floors to ground
• Remaining proportion of people going between the remaining floors are split equally

Future Works

• Update algorithm of elevator movement. Current one has some flaws in logic
• Also create elevator that has capacity constraints, currently there are no limitations to capacity