Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!lll-winken!uunet!ncrlnk!ncrcae!hubcap!agn From: agn@unh.cs.cmu.edu (Andreas Nowatzyk) Newsgroups: comp.parallel Subject: Re: Looking for Petri net 'Toolkit' Keywords: petri net, modeling Message-ID: <5049@hubcap.clemson.edu> Date: 10 Apr 89 12:16:08 GMT Sender: fpst@hubcap.clemson.edu Lines: 48 Approved: parallel@hubcap.clemson.edu I used "GreatSPN" written by Giovanni Chiola which is based on M. Molloy's work on generalized Petri Nets. It is a very nice package: - Interactive graphics editor (run on SUNs using Suntools) that produces internal representations and postscript output. - The editor knows about PNs and does sanity checks. - Deterministic, immediate and exponentially timed transitions - Inhibitor arcs are supported. - Analysis tools can be called from the editor and backannotate into the graphical representation. (Several analytical analysis modules are included plus a Monte Carlo simulator for cases that exceed the analytical capabilities) - Error bounds on the results are given. Results appear as bar-graphs or in numeric form. - Interactive net execution (to debug the net) - Static analysis tools are provided to check boundedness etc. Minor drawback: the simulator was written in pascal and was quite slow. I rewrote that in C (left out a few features that I didn't need plus added others: about 50-100x speedup, or about 100K transition fires/sec on a SUN 3/60. Nets with 500 places and 1000 transitions were successfully solved. My code is free for the asking, GreatSPN requires permission from the author, who can be reached on BIT-NET: CHIOLA%ITOINFO.BITNET%IBOINFN.BITNET I can redistribute GreatSPN if evidence of the author's permission is presented. This may speed up things (Smail round-trip for a tape to Italy isn't great), but leaves you with version 1.4 which may not be the most recent. -- -- Andreas Nowatzyk (DC5ZV) Carnegie-Mellon University agn@unh.cs.cmu.edu Computer Science Department (412) 268-3617