compleathorseplayer

Thanks. I am used to DE/PSO, where the default is to keep the previous best in the population and usually the current best as well. I don't have good intuition...

Hi. I managed to get it working with the improved parameters. I think there must be something wrong with the implementation, because the objective function keeps on getting bigger and...

Hi. Thanks, I will try examples like that myself, and compare with BlackBoxOptim Is there any place where examples like the above are shown? It is very difficult to guess...

Hi. Has anyone apart from you gotten this to work? I have used GA and Evolutionary Optimisation in the past (and even teach courses on it) so I have some...

Hi Art, My main suggestion would be to make the defaults more reasonable. Given how bad my results have been, I also wonder whether the default tolerance is in the...

Hi Art, First, I verified that I could run your regression example using CMAES, and that worked fine. Next, I tried your Suggestion above for the GA. There were a...

Hi Art, Thanks for getting back to me. I will re-download and try the CMA-ES. In terms of the GA, I have worked out my toy problem. I think you...

Thanks so much for taking a look! Yes, that is the solution! Of course I know the necessity for binary crossover but did not understand the syntax of your previous...

That's great. Regarding your comment about the API, I wonder if I might suggest that if (as it seems) the algorithm falls over for bit string arguments if the operations...

Thanks. Students seem to suddenly have this problem even on (ISLR) datasets which have previously worked fine with exactly the same syntax