From Ecological Modeling to Decentralized Optimization of Smart Power Grids

In this contribution we want to sum up how computer science, ecological modeling and control of distributed power systems mutually influence. First we show, how ecological modeling has been influenced by the availability of faster computers and easier to use programming environments. Then we argue that adaption and emergent behavior in biological systems observed e.g. at the hand of microscopic models can be interpreted as an optimization process. This observation led to new optimization (meta-)heuristics in computer science and inspired multi-agent based optimization and control methods. We motivate why these methods are an interesting choice for finding new solutions to planning and control problems in future electrical power systems. As an example we sketch a completely distributed, agent-based planning and control system for a so-called dynamic virtual power plant. Finally we discuss some directions of future research in the domain of agent-based control of distributed systems.