Hey everyone, in the last weeks I've been thinking about building a power market simulation engine that is not based on any linear or mixed-integer programming paradigm because these are hopeless when faced with making decisions under uncertainty. I've built a demonstratotr using julias out of the box genetic algorithm and I am quite fascinated by how far one can push this.I've modeled europes core powermarkets,connected via and PTDF approximation, fed with real-ish data. Number of plants is still at least an order of magnitude short of any real number, but in principle this scales quite well because the algorithm can be parallelized.

I'd be interested if anyone knows of any production-level systems that use this approach or any feedback in general. Or any other intel about howto incorporate stochasticity in energy simulation.