From: Nick C. <nic...@ve...> - 2005-09-13 12:34:26
|
YOUNG (I think you mean JUNG here though) could well be faster. It's certainly worth trying. Converting the network shouldn't be much work at all. You are free to use the NTRecorder code from repast in anyway you like. Certainly feel free to do so. Nick -----Original Message----- From: rep...@li... [mailto:rep...@li...]On Behalf Of Taymaz Khatami Sent: Tuesday, September 13, 2005 4:42 AM To: rep...@li... Subject: [Repast-interest] NT Recorder and Average Path Length HI again. Today I want to post two problems I'm dealing with. 1.. Is there any other way to calculate the Average Path Length, then the method of NetUtilities? As I mentioned earlier this method takes a lot of computation time (1/2 day on a 1000 Node matrix). Has anyone an idea if the method of YOUNG is in some way more efficient? And if so how I could integrate it, because the YOUNG's method works on YOUNG specific graphs and in case I use that method, I had to convert my agentList to a graph (or ArchtypeGraph) first before calling the method. I guess that would work this way but I'm not sure. 2.. The NTRecorder of Repast produces a consecutive DL-File which gets pretty huge when running 5000 simulations (approximately 10 GB). Could I hypothetically "steal" the source code from the Repast Library (of course with mentioning the Source and the authors) and fit it to my purpose, or are there any copy right violations. Actually I need a NTRecorder method, which creates me like 5000 separate DL-files and deletes the header, as UCI Net seems not beening able to read the file with that header in it. Thx guys |