Talk SimPy, a discrete event simulation package in Python

Presented by Klaus Müller in Scientific Applications 2008 on 2008/07/26 from 10:00 to 11:15
Abstract
The presentation will address SimPy (Simulation in Python), an Open Source discrete event simulation package. It is written completely in Python. SimPy is released under a LGPL license and can be downloaded free of charge from http://sourceforge.net/project/showfiles.php?group_id=62366. The SimPy website http://simpy.sourceforge.net describes the package in detail and provides extensive user documentation. SimPy is being used for teaching and research purposes at universities in several countries. Many companies and institutes also use it for simulation studies. The talk describes the basic philosophy and design of SimPy. SimPy is modeled after the Simula67 simulation language, but goes beyond that language’s capabilities in terms of simulation constructs and extensibility. It is a process-oriented simulation package, i.e., the lifecycle of a simulation entity and all its events are described in one method (called the Process Execution Method). Processes are implemented as Python generators. Process scheduling is done by yield statements with parameters. This provides semi-coroutines, light-weight threads which run in pseudo-concurrency. Various small examples of SimPy programs will be presented and run to show the range of process control constructs, and also to illustrate how concisely and descriptively models can be programmed in SimPy. The various run-modes of SimPy (which include event tracing, event-by-event execution, and synchronization between wall-clock and simulation time) will be described and run, as will be the GUI and plotting libraries included. The talk will conclude with a few words about the relationship between SimPy and SciPy.
tagged by
no related entity