Feedforward Neural Network (NN) model with difference-of-convex-functions (DC) structure.

Table of Contents

1. About The Project
   • Built With
2. Getting Started
   • Prerequisites
   • Running the code
   • Options
3. Application
4. Contact

About The Project

Learn dynamical systems from data using feedforward Neural Network (NN) models with a special difference-of-convex-functions (DC) structure. The so-called DC-NN model approximates the system dynamics f(x, u) in DC form as follows: f = f1 - f2 where f1, f2 are convex functions with respect to (x,u). The NN model structure is such that the functions f1, f2 are approximated independently by two input-convex NN with their own sets of weights. The DC-NN model is leveraged in this example to predict the dynamics of the
coupled tank system.

Built With

• Python 3
• Keras

Getting Started

Prerequisites

You need to install the following:

• numpy
• matplotlib
• tensorflow / keras

Run the following command to install all modules at once

pip3 install numpy matplotlib tensorflow

Running the code

1. Clone the repository

   git clone https://github.com/martindoff/DC-Deep-Neural-Network.git

2. Go to directory

   cd DC-Deep-Neural-Network

3. Run the program

   python3 DC_NN_model.py

Options

1. To load an existing model, set the load variable in DC_NN_model.py to True

   load = True

Set the variable to False if the model has to be (re)trained.

Application

Such model have applications, e.g. in the framework of robust tube MPC for systems representable as a difference of convex functions (see paper on DC-TMPC) The DC-NN model allows one to learn the dynamics of any sufficiently regular system in DC form and then apply the DC-TMPC algorithm.

Contact

Martin Doff-Sotta - [email protected]

Linkedin: https://www.linkedin.com/in/mdoffsotta/