v0.2.0

v0.2.0 (2019-06-12)

• Breaking changes: rename NoOptimizer -> NoModelOptimizer, MLGPOptimizer -> MAPGPOptimizer, lhs_iterations -> initializer_iterations
• New features: use Sobol sequence by default to initialize the model. Latin hypercube sampling can be used with ScaledLHSIterator.
• Fixes: warmstart models (#7, #10)

Diff since v0.1.1

Closed issues:

• Restarting optimization (#7)
• Ambiguous error message for incorrect type (#6)
• Allow GP to use given observation noise (#5)

Merged pull requests:

• Warmstart (#7) (#10) (platawiec)
• correct a trivial typo (#4) (tpapp)