Release Notes

Some notes on new features in various releases

What’s new in 0.5.2

API Changes:

  • Add support for Python 3.12, 3.13, and 3.14.

  • Update GitHub Actions tests on all supported Python versions.

  • Update project dependencies to versions compatible with modern Python and NumPy releases.

  • Update ParameterGrid to work with NumPy 2.x.

  • Validate Erlang C systems with zero productive positions or traffic intensity greater than or equal to positions.

  • Validate required_positions() with max_occupancy > 0.

  • Fix the MultiErlangC inconsistent-results error message to report the expected number of scenario combinations.

What’s new in 0.5.1

Features:

  • Remove support for Python 3.7 and add support for Python up to 3.11.

  • Update the project dependencies

What’s new in 0.5.0

Features:

  • Added a new type of solver under the class MinHoursRoster for rostering problems, it can find the roster of resources for each day and shift subject to shift restrictions, resting days, shifts preferences, bans, and more.

  • Added the properties waiting_probability_params, service_level_params, achieved_occupancy_params, and required_positions_params in MultiErlangC to track in which combination order each method returns a solution.

API Changes:

  • The misspelled queing module was renamed to queuing.

  • The shifts module was renamed to scheduling.

What’s new in 0.4.1 and below

  • Implemented ErlangC for solving queue systems positions requirements

  • Implemented MultiErlangC as a parallel implementation for multi-input ErlangC, similar to scikit-learn’s param_grid in Grid Search

  • Added MinAbsDifference and MinRequiredResources solvers to find the optimal number of resources to allocate in a shift, based on a pre-defined requirement of the number of resources per period of the day.

  • GitHub Actions for pytest and Codecov reports

  • Examples and tutorials on all the package features

  • Initial docs setup