neural-body 0.1.1

An n-body simulator

Neural Body N-Body Simulator

An n-body simulator powered by a neural network.

Neural Body is an n-body simulator currently as an alpha demonstration of substituting calculations for planetary motion with a neural network. Currently, the neural networks can predict the position of Mars or Pluto given the rest of the positions in the planetary system.

The eventual goal is to replace the physics simulator that calculates the positions of other bodies to feed the neural network completely with neural networks that take in an acceleration vector on each body along with a desired simulation time step and perform the integration necessary to calculate the displacement of the body.

Considerable time was put into training the first version of the neural network which was a simple, 2 layer, feedforward neural network with 300 nodes per layer that performed multiple-output regression on the x, y, z coordinates of the body being predicted. Input to the network was the position of every other body in the system at each time step of a previously run simulation.

Below is a Google Colab notebook that shows the output of a training run for the neural network that predicts Mars' position. Code that generated the training data and performed preprocessing is not included. Data file is also not included. The link is purely to view code, learning curves, and results.

Mars Neural Network Training

---

Table of Contents

• Installation
• Usage
• Documentation
• Team
• License

Installation

Requirements

• Compatible debian-based Linux distro. Ubuntu 20.04 Preferred.
• Python 3.8 or higher.
• PyGame 2.0.0.dev10 or higher.
• TensorFlow 2.2.0
• Pandas 1.0.5
• Numpy 1.19.0 All dependencies above except for Python 3.8 should install when pip install is run.

Setup

• Download neural_body-0.1.0.tar.gz from the dist directory.
• Navigate to local folder where download is located.
• Use pip install neural_body
• Use the neural_body command to run the simulator.

---

Usage

This selection includes an overview of all menu buttons and functionality of the simulator.

Pause / Play

The simulation can be paused at any point.

Toggle View

The simulation view can be toggled from overhead to side view.

Adjust Speed

The simulation can be sped up or slowed down.

New Simulation

The initial state of all bodies in the system are contained in CSV files packaged with the simulator. Whichever planet is designated as the "satellite" tells the simulator which neural network to use for planetary motion prediction. This early demo can only predict the motion of Mars or Pluto given the positions of the other planets in the system. In later releases, the neural network will be updated to accommodate predicting the motion of any body.

Current Config File Options:

• mars_sim_config.csv
• pluto_sim_config.csv

Is NASA Right?

If you disagree with NASA, you can bring Pluto back as a planet.

Show Planet Key

Hovering over this option displays a color coded key of all planets in the system.

Travel to a Day

Selecting this option allows the user to rewind or fast forward the simulation by entering the day they would like to jump to. There is a heavy delay for fast forwarding as the simulator right now must inefficiently calculate every frame between the current day and the day you entered. Negative time values will be treated as reverting back to 0 day.

Documentation

To view the documentation online, go to the following URL:
Neural Body Sphinx Documentation

Documentation for the source compiled with Sphinx is included in the neural-body/0_demo-sim/docs/_build/html/ folder. You will need to download and host yourself by running python3 -m http.server --directory _build/html

Team

The AstroGators formed as a result of the "CIS4930 - Performant Python Programming" course at the University of Florida.

Team members include:

• Nathaniel Edgar
• Craig Boger
• Gary Jones
• Cory Robertson
• Andrew Sowinski

The Github repo for this initial demonstration is located at: https://github.com/nedgar76/neural-body/tree/demo-sim/0_demo-sim

License

• MIT license
• Copyright 2020 ©