SimPy and Dynamic DES

Dynamic DES is built on SimPy, a Python library for discrete-event simulation. SimPy provides the event list, the clock, processes and resources. Dynamic DES keeps all of them and changes four things: the clock can switch pacing, every parameter can change while the run goes on, random times are named distributions, and everything that happens is published as records. A low-level Dynamic DES model is ordinary SimPy code with these additions.

Part	In SimPy	In Dynamic DES
Clock and pacing	Environment runs without waiting; RealtimeEnvironment(factor, strict=True) follows the wall clock and raises RuntimeError when the model falls behind	one class, DynamicRealtimeEnvironment, does both: factor=0 runs without waiting, go_live_at switches to real time partway through a run (SimPy's factor cannot change once set), logical_start_time gives every record a date and time, and falling behind is reported as telemetry rather than raised
Processes	generators started with env.process	the same, and the builder adds @app.task, @app.arrival_loop, @app.telemetry_loop and add_process, which write the common ones for you
Random times	none built in: a model calls random or NumPy itself	five named distributions with checked parameters, drawn by one seeded Sampler from the current values each time
Resources	Resource, Container and Store keep the capacity they were created with: it has no setter, so assigning a new one raises AttributeError (SimPy 4.1.1)	DynamicResource, DynamicContainer and DynamicStore follow a capacity in the registry, and a resource that shrinks lets busy units finish
Parameters	Python variables, changed only by the model's own code	a registry of named paths, changed by connectors, scenarios or processes, with each change checked
Output	whatever the model prints or logs	events and telemetry with a fixed shape, written to Kafka, Redis, PostgreSQL, files or Iceberg on background threads

Each building block page has a section on what SimPy does for that block and what Dynamic DES changes.