v2.0.1

This release deprecates DRS. Subsequent research has found that the Dirichlet-Rescale algorithm does not always return uniform values. Unless studying the effect of this non-uniformity, it is not recommended to use DRS.

For an algorithm which provides a correct uniform sampler for the same problem, please use the ConvolutionalFixedSum algorithm. The associated research also includes information on how the non-uniformity present in DRS can be triggered.

v2.0.0

Changes to the behaviour of the algorithm mean that this release breaks compatibility with v1.0.0.

• Changes behaviour so that lower_bound == upper_bound behaves as expected rather than being an error
• Replaces constraints with bounds (shorter variable names as before these were getting out of hand)
• Updates internal variable names to be more consistent and less influenced by the real-time systems use case
• Improvements to documentation

v1.0.0

Initial release of DRS, corresponds to version used for publication. Future versions will add new features / fix bugs.