Introduction

MeLOn provides scripts for the training of various machine-learning models and their C++ implementation which can be used in the open-source solver MAiNGO. The machine-learning module git repository currently contains the following models:

Feedforward neural network for regression
Gaussian process for regression (Kriging)
Multifidelity Gaussian process for regression
Support vector machine for regression
Support vector machine for outlier detection (one class support vector machine)
Convex hull of point cloud

Note that MeLOn is a submodule of our global solver MAiNGO. Thus, it will be automatically included as a dependency when you download MAiNGO and initialize & update the submodules. Then, the developed machine learning models can be used within MAiNGO.

Also, note that we have developed various envelopes or tight convex and concave relaxatoins of relevant functions. These functions are implemented within Mc++. When using MeLOn or MAiNGO, Mc++ is automatically inlcuded as a submodule.

This manual is divided in the following sections: