Possible typo in population size of first regime

[vinaym815]

libcmaes/src/bipopcmastrategy.cc

Line 122 in bf99435

double ltmp = pow(0.5*(_lambda_l/_lambda_def),u);

Should not the value sampled from uniform distribution be squared as suggested in:
https://hal.inria.fr/inria-00382093/document
https://hal.inria.fr/hal-00818596v1/document

Also it might be desirable to change the sigma expression to the one proposed in second paper.
sigma expression of first publication:

sigma expression of the second publication:

[beniz]

Hi @vinaym815 good catch, possibly, let's see whether @nikohansen can confirm ?

[nikohansen]

Yes, I agree with both, 2 -> default initial sigma and the uniform random number should be squared first.