rewriting_my_dde_solver

In the final months of writing my master thesis there was little time. Simultaneously a lot of problems in the c++ code that I had written became very apparent. Since then it has been a dream of me to rewrite the whole thing. Thereby learning even more C++ and also using git/github in a more professional manner (I used it before, but only locally and I didn’t quite adhere to a consistent way of commiting changes or wording the commit messages).

My rewrite can be found here on my github. The goal will be to reimplement the features of the old solver, but in a more modular and flexible manner. The features of the old DDE solver were:

• construction of networks as a collection of edges with individual coupling weights, phases and delays
• implementation of the Stuart-Landau system
• integration of the delay-coupled systems with runge-kutta 4th order method
• analysis of the timeseries - finding extrema
• saving timeseries in a simplified manner to dramatically save space [RDP]
• measuring (non-)linear reservoir computing performances via Legendre-Polynomials [DAM12]

Additional features that shall be in the new solver:

• adaptive refinement of the parameter space when scanning
• implementation and easy changing of advanced dynamical systems e.g. Lang-Kobayashi, Lorenz-sytem, Rössler-System and more.

[RDP] using the Ramer-Douglas-Peucker algorithm

[DAM12] Dambre, J., Verstraeten, D., Schrauwen, B. et al. Information Processing Capacity of Dynamical Systems. Sci Rep 2, 514 (2012). https://doi.org/10.1038/srep00514