Model Aggregation

This was our 5th semester project. It is about using optimizations techniques to find lumping matrixes for systems of ODEs. For more details please see report.pdf. In this project we used gradient descent and particle swarm optimization. Within gradient descent we explored different techniques for determining the learning step: golden section search, quadratic interpolation optimization, RMSProp, ADAM, MOMENTUM and modified dynamically scaled versions of the last 3.