isaaccheng9 / lift-control-simulation

A desktop application to simulate lift control with a custom algorithm optimised to reduce distance travelled by the lift by 82.5%. Developed with Qt.

Lift Control Simulation

A desktop application developed in Python with PyQt6 to manage and simulate lift control under various configurations.

I optimised a custom algorithm for efficiency in distance travelled by the lift. This algorithm achieves up to 82.5% reduction in distance travelled compared to the naive algorithm benchmark, outperforming the 25% reduction claimed by the destination dispatch technique.

For more information about this, see the design document at docs/Lift Control - Design Document.pdf. It contains an explanation of the algorithm and performance analysis.

Visuals

Installation

Operating System

Compatible with:

• Windows
• MacOS
• Linux

Python Version

Python 3.7.0 or later is required for this software to run. It can be downloaded from Python's website here.

Running the Application

To run the application, you should follow the following steps:

1. Clone this GitHub repository.
2. Ensure that you're in the root directory: lift-control-simulation
3. Install the required Python libraries: pip install -r requirements.txt
4. Run the application with the command: python -m src.app

Usage

Configuration

The user is able to configure any number of floors, number of people, lift capacity, and the delay of intervals in the user interface.

Simulation

Their simulation can be visually displayed for up to five floors. For more than five floors, the application provides a non-visual display. In both cases, the user can run their simulation with either the naive mechanical lift (the benchmark) algorithm, or my improved lift algorithm.

The configured simulation remains the same until the user generates a new simulation, or changes the configuration settings. This enables the user to directly compare how the improved algorithm compares to the naive algorithm.

Logging

The application prints a temporary log of the simulations in the terminal. It records the simulation configuration, people generated, movements of the lift, people pending (if applicable), people in the lift, people delivered, and an overview of people after the simulation is complete.

lift-control-simulation's Issues

Create auxiliary functions for repeated code

This would improve:

• Readability to distinguish between different lift control algorithms more clearly
• Modularity for future implementations of new lift control algorithms

Add option to perform a set of benchmark simulations

This would make the differences in performance between the naive algorithm and the improved algorithm clearer to the user. It would perform five sets of simulations using both algorithms, and then show the results in a new window. The new window would display:

• A table of results
• A graph to compare results visually (Matplotlib?)

Migrate GUI from PyQt5 to PyQt6

This will improve the handling of high DPI scaling, which is an increasingly common problem due to resolutions in laptops and monitors increasing.

Display the difference in distance travelled between naive and improved algorithms

This would help to demonstrate how much more efficient the improved algorithm is.

Create people_overview.json if it doesn't already exist

This would prevent us from needing to track a people_overview.json file for easier maintenance and less clutter.