+ PhaseType distributions implemented + Classes for specific PhaseType distributions: Erlang, HyperExponential, HyperErlang, and Coxian
2.2.0
+ Processor sharing implemented (limited and capacitated) + Ability to use a different node_class per node of the network + State tracking now works with simulate_until_max_customers + Remove testing on Python 3.5
2.1.3
+ Small refactor to Node adding new servers, and to Individuals receiving the Simulation object. + Add a library of custom behaviour to docs + Support Python 3.8, update hypothesis
2.1.2
+ Ability to incorporate customer behaviour Server and Individual classes.
2.1.1
+ State Trackers slightly more efficent, they do not record any state changes that result in the same state as before. + Add the NodePopulationSubset tracker. + Distribution objects can now see the Simulation object, for true state dependent distributions.
2.1.0
+ State Trackers now track history + State Trackers give state probabilities + A number of performance improvements + Fix some documentation + Test on PyPy3.6 and Python 3.7 too