1Blurb:: Portion of batch size dedicated to exploration 2 3Description:: Refinement candidates are generated by an acquisition 4function such as maximum expected improvement, which balances 5exploration and exploitation. Refinement candidates can also be 6generated by purely explorative metrics such as maximum prediction 7variance. For a specified \c batch_size, \c exploration specifies the 8subset of this total that will be dedicated to pure exploration of the 9parameter space. 10 11<b> Default Behavior </b> 12All of the batch size is devoted to the standard acquisition approach, 13balancing exploration and exploitation . 14 15Topics:: 16 17Examples:: 18\verbatim 19method, 20 efficient_global 21 seed = 1237 22 batch_size = 8 # total 23 exploration = 2 # 2 out of 8 24\endverbatim 25 26Theory:: 27Faq:: 28See_Also:: 29