Given examples work better using Differential Evolution

[dietmarwo]

Thank you for providing these interesting optimization examples. I adapted them for use with Python3 and added them as examples to my optimization library referencing this repo as original source:

https://github.com/dietmarwo/fast-cma-es/blob/master/examples/robot.py and https://github.com/dietmarwo/fast-cma-es/blob/master/examples/rover.py and documented them here https://github.com/dietmarwo/fast-cma-es/blob/master/tutorials/RobotRover.adoc.

Suprizingly it seems both work very well using Differential Evolution. Only a few seconds are required to surpass the EDO results from https://arxiv.org/pdf/1706.01445.pdf. Do you have a real world example where the huge computational overhead of EDO really makes sense?

[qlchen2117]

For sampling-efficient algorithms, the number of evaluations is more important than the running time. As is known, Evolutionary algorithms have low time complexity. I think you could add more experiments about regrets with respect to evaluations.