http://permalink.gmane.org/gmane.comp.python.simpy.user hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://permalink.gmane.org/gmane.comp.python.simpy.user/1360 <pre>Hello, I am trying to simulate a production line, all the way from customer order to dispatch. This will allow me to load the factory with different types of work based on historic data. But a down side of this, I need to simulate the process of order creation. So first the customer places and order, that order has a number of lines and the line has a number of units. I have made a procedural source which can create the right number of number lines per order and units per line based on some variables. But the problem I am facing is; the factory stores kits a parts for an entire order line (based on which parts they have), but then the builders pick individual units (based on the units which are kitted). So I need to be able to manipulate an entire line at one stage, but at a later stage I need to work with individual units. This seems like the sort of thing that some one might have done before (possibly using OO), so is there an example of this structure that I could follow? class Source(</pre> Chris 2012-05-23T14:08:01 http://permalink.gmane.org/gmane.comp.python.simpy.user/1359 <pre>So essentially, in principle, we CAN assume that the units are chosen 'at random' since it would not make any difference to the results in any way? On Tue, May 22, 2012 at 9:39 PM, Tony Vignaux &lt;vignaux&lt; at &gt;gmail.com&gt; wrote: </pre> Josephine Lim 2012-05-22T09:58:46 http://permalink.gmane.org/gmane.comp.python.simpy.user/1358 <pre>Precisely. The number of "free" units is just an integer, a count. The units are not distinguished at all. If you need to control particular units you will need to program this yourself. For example, you could have a Resource(2,..) controlling access to the units then have, some programming that chose between particular units. ==== Tony Vignaux On Tue, May 22, 2012 at 8:42 PM, Josephine Lim &lt;josephinelim86&lt; at &gt;gmail.com&gt;wrote: ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/_______________________________________________ Simpy-users mailing list Simpy-users&lt; at &gt;lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/simpy-users </pre> Tony Vignaux 2012-05-22T09:39:57 http://permalink.gmane.org/gmane.comp.python.simpy.user/1357 <pre>Actually, now that I think about it, it does seem to allocate 'randomly', in the sense that it does not distinguish in any way or form between multiple units of resources. So to SimPy, it doesn't matter which unit is used at all, they are not even viewed as separate objects. Would that be accurate? On Tue, May 22, 2012 at 7:12 PM, Stefan Scherfke &lt;stefan&lt; at &gt;sofa-rockers.org&gt;wrote: </pre> Josephine Lim 2012-05-22T08:42:11 http://permalink.gmane.org/gmane.comp.python.simpy.user/1356 <pre> Am 2012-05-22 um 09:06 schrieb Josephine Lim: This is not possible with SimPy’s built-in resources. Maybe you can inherit them do implement your required behaviour. Cheers, Stefan ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ </pre> Stefan Scherfke 2012-05-22T07:12:41 http://permalink.gmane.org/gmane.comp.python.simpy.user/1355 <pre>Hi Stefan, I re-read your example, and I think I need to clarify what I mean. In your example, you mean that if B and C request a resource at the same time B will always get it, right? Because you have not specified priorityQ as your resource parameter, so that will be the default case. I understand that. What I am asking is if I have two units of a resource, eg. k = Resource(capacity=2, name="Printer", qType=PriorityQ) . Both units of resource are free at a given time. Now A requests a resource. I want there to be a 0.5 chance of the first resource unit being allocated to A, and a 0.5 chance of the second resource unit being allocated to A. ie if there are two printers available, A will choose one of them at random for his task. Thanks again. On Tue, May 22, 2012 at 6:50 PM, Josephine Lim &lt;josephinelim86&lt; at &gt;gmail.com&gt;wrote: </pre> Josephine Lim 2012-05-22T07:06:22 http://permalink.gmane.org/gmane.comp.python.simpy.user/1354 <pre>Is there a way that I can alter that to induce randomness with a probability of 0.5 for A and B (I don't require a C)? On Tue, May 22, 2012 at 6:48 PM, Stefan Scherfke &lt;stefan&lt; at &gt;sofa-rockers.org&gt;wrote: </pre> Josephine Lim 2012-05-22T06:50:12 http://permalink.gmane.org/gmane.comp.python.simpy.user/1353 <pre>Hi Josephine, there is no randomeness involved. In the following example, B will get the resource always before C: from simpy.Simulation import Simulation, Process, Resource, hold, request, \ release class TestProc(Process): def __init__(self, sim, name, resource, init_hold=0): super().__init__(sim=sim, name=name) self.resource = resource self.init_hold = init_hold def run(self): yield hold, self, self.init_hold print('%s requesting resource at %s' % (self.name, self.sim.now())) yield request, self, self.resource print('%s got resource at %s' % (self.name, self.sim.now())) yield hold, self, 4 print('%s releasing resource at %s' % (self.name, self.sim.now())) yield release, self, self.resource sim = Simulation() resource = Resource(sim=sim, capacity=1) a = TestProc(sim, 'A', resource) b = TestProc(sim, 'B', resource, 2) c = TestProc(sim, 'C', resource, 2) </pre> Stefan Scherfke 2012-05-22T06:48:05 http://permalink.gmane.org/gmane.comp.python.simpy.user/1352 <pre>Lis ton message de Syn avant qu'il ne soit effacé! Pour lire ton message, suis simplement ce lien: http://eu1.badoo.com/0279183266/in/hbFeqxWJOkU/?lang_id=6&amp;m=65&amp;mid=4fbb0ff60000000000060000610b9a0b Le lien ne fonctionnent pas dans ce message ? Copie le dans la barre d'adresse de ton navigateur. Tu as reçu cet email suite à une requête de Syn auprès de notre système. S'il s'agit d'une erreur, ignore simplement cet email. La requête sera alors effacée du système. Amuse-toi bien ! L'équipe Badoo Vous avez reçu cet email de Badoo Trading Limited (adresse postale ci-dessous). http://eu1.badoo.com/impersonation.phtml?lang_id=6&amp;mail_code=65&amp;email=simpy-users%40lists.sourceforge.net&amp;secret=&amp;action=block&amp;block_code=3b37f6&amp;m=65&amp;mid=4fbb0ff60000000000060000610b9a0b Badoo Trading Limited est une Société à Responsabilité Limitée, enregistrée en Angleterre et Pays de Galles, sous le No. 7540255, bureaux enregistrés à l’adresse 12 Red Lion Square, London, WC1R 4QD.-----------------------------</pre> Badoo 2012-05-22T04:03:04 http://permalink.gmane.org/gmane.comp.python.simpy.user/1351 <pre>Hi, Sorry to bring this up at such a late date, but I looked into the Resource() class of SimPy and could not find an indication of how it chooses a unit to be used when both units of resource are empty and 1 job is in queue. Is there any way I can prove that a unit is chosen randomly? Thanks again. On Thu, May 17, 2012 at 12:20 AM, Josephine Lim &lt;josephinelim86&lt; at &gt;gmail.com&gt;wrote: </pre> Josephine Lim 2012-05-21T22:31:09 http://permalink.gmane.org/gmane.comp.python.simpy.user/1350 <pre>Hi Linus, Am Montag, 21. Mai 2012, 19:20:49 schrieb Linus Junden: Indeed, I was puzzled too :) Interesting, your program doesn't use more than 38mb without the __del__ methods. Here's the output: $ python testdelete.py ... let's look at how SimPy Simulation classes work... 0 all simulations still stored in memory... 100000 all simulations still stored in memory... 200000 all simulations still stored in memory... 300000 all simulations still stored in memory... 400000 all simulations still stored in memory... 500000 all simulations still stored in memory... 600000 all simulations still stored in memory... 700000 all simulations still stored in memory... 800000 all simulations still stored in memory... 900000 all simulations still stored in memory... obviously simpy stores references to all simulation instances created... deletes B number 9 from memory This is with 2gb ram: $ cat /proc/meminfo MemTotal: 2018972 kB What's your python versi</pre> Ontje Lünsdorf 2012-05-21T18:27:04 http://permalink.gmane.org/gmane.comp.python.simpy.user/1349 <pre>Well if it's not a bug in SimPy I would at least say that it's an unfortunate property. you are wrong about the del thing, the example code crashes because of lack of memory without the __del__ methods as well (if it doesn't you have more ram than me, add a factor of 10 to the simulations and try). Thanks for looking into it though. On 18 maj 2012, at 08:45, Ontje Lünsdorf &lt;the_com&lt; at &gt;gmx.de&gt; wrote: ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Simpy-users mailing list Simpy-users&lt; at &gt;lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/simpy-users </pre> Linus Junden 2012-05-21T17:20:49 http://permalink.gmane.org/gmane.comp.python.simpy.user/1348 <pre>Hi Seymour, Am Donnerstag, 17. Mai 2012, 14:09:12 schrieb Seymour Morris: Yes, this was the problem, you fixed it yourself :). In your original version you used the global br: This br was always the last Breakdown() instance created in the loop. On a side note: I actually don't know why br needs to be reactivated. It is not passivated at any point in time. I think there's a bug in the example? Cheers, Ontje ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ </pre> Ontje Lünsdorf 2012-05-18T07:01:53 http://permalink.gmane.org/gmane.comp.python.simpy.user/1347 <pre>Hi Linus, Am Mittwoch, 16. Mai 2012, 18:02:25 schrieb Linus Jundén: This isn't a bug in SimPy. Remove the __del__ methods from testProcess and simTest and the simulations will be properly garbage collected. This is a garbage collection pecularity. Objects in a reference circle and with a __del__ method are not automatically garbage collected because python cannot guess the order in which those object should be garbage collected. See http://docs.python.org/library/gc.html#gc.garbage for more details. Cheers, Ontje ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ </pre> Ontje Lünsdorf 2012-05-18T06:45:30 http://permalink.gmane.org/gmane.comp.python.simpy.user/1346 <pre>Hi Linus, Am 2012-05-16 um 18:02 schrieb Linus Jundén: Sounds very interesting. :-) SimPy 3 may support objects with multiple PEMs. Maybe Python’s garbage collector has problems with the circular references between processes and simulations. You could try to manually delete your Process and the simulation objects (if you are not using the global Simulation instance. We’ll keep that problem in mind for SimPy 3. Best regards, Stefan ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ </pre> Stefan Scherfke 2012-05-17T19:58:14 http://permalink.gmane.org/gmane.comp.python.simpy.user/1345 <pre> Stefan Scherfke &lt;stefan &lt;at&gt; sofa-rockers.org&gt; writes: Have you tried to step through the Hi Stefan, Thanks for the reply. I did experiment with SimulationTrace and still could not figure out why the last instance was getting "off sequence" using this approach and I do not currently understand pdb well enough to use it. After much more trial and error experimentation I found that if I define an __init__ method within the Bus object and activate the Breakdown process at this point, then all works as expected. Here is the revised code simulating four Buses that fail exactly on 300 hour intervals and are repaired in exactly 20 hours: from SimPy.Simulation import * class Bus(Process): def __init__(self,name): Process.__init__(self,name) self.name = name self.br=Breakdown(self) activate(self.br,self.br.breakBus(300)) def operate(self,repairduration,triplength): # PEM tripleft = triplength while tripleft &gt; 0: yield hold,self,tripleft # try to finish the trip</pre> Seymour Morris 2012-05-17T14:09:12 http://permalink.gmane.org/gmane.comp.python.simpy.user/1344 <pre>Hi Seymour, this looks strange. At a first glance, I can’t see a problem in your code. Have you tried to step through the simulation using the Python Debugger (pdb) or via a SimulationTrace object? Cheers, Stefan Am 2012-05-16 um 23:40 schrieb Seymour Morris: ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ </pre> Stefan Scherfke 2012-05-17T10:13:49 http://permalink.gmane.org/gmane.comp.python.simpy.user/1343 <pre>Hello, I have been experimenting with the Bus breakdown example in the Simpy manual and am really struggling to understand why when I create multiple instances of Bus, the last instance seems to get "of sequence" after the first repair. I have modified the example code in the manual very slightly just below the initialize() statement to create two instances of Bus (Bus1 and Bus2). Here is my code: from SimPy.Simulation import * class Bus(Process): def operate(self,repairduration,triplength): # PEM tripleft = triplength # "tripleft" is the driving time to finish trip # if there are no further breakdowns while tripleft &gt; 0: yield hold,self,tripleft # try to finish the trip # if a breakdown intervenes if self.interrupted(): print self.interruptCause.name, 'at %s' %now() tripleft=self.interruptLeft # update driving time to finish # the trip if no more breakdowns self.inte</pre> Seymour Morris 2012-05-16T21:40:15 http://permalink.gmane.org/gmane.comp.python.simpy.user/1342 <pre>Not trying to fix the problem, but you can use multiprocessing module to run simulations in separate processes. This will not make problem disappear, but it won't cause problems since only one (or limited number) simulation is done per process. And you get concurrency also. On Wed, May 16, 2012 at 7:02 PM, Linus Jundén &lt;linus.junden&lt; at &gt;gmail.com&gt; wrote: ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ </pre> Harri Vartiainen 2012-05-16T16:31:39 http://permalink.gmane.org/gmane.comp.python.simpy.user/1341 <pre>Dear User-community I have been using SimPy for over a year now in a forest machine simulation project. It's complex simulations with a lot of data involved connected to the spatial complexity of a forest. See http://www.org.umu.se/umit/english/project-activities/discrete-event-forestry/ for more info about the project. Overall I like SimPy although we have problems with the one-PEM limitation which unfortunately have resulted in close to unreadable code. But enough about that. Recently we had big simulation series with around 50 000 simulations. Simply put, we ran out of memory. It turns out that all the 50 000 simulations are stored in memory and not garbage collected. And this is NOT because we still have references to those simulations from our code. It turns out that SimPy stores references to earlier simulations. Try running the below/attached example and you will see what I mean. I haven't time right now to look into the SimPy code to see what's wrong... do you have any hints? Best Regards Linus</pre> Linus Jundén 2012-05-16T16:02:25 http://permalink.gmane.org/gmane.comp.python.simpy.user/1340 <pre>Hi Tony, No, if the units of resources are randomly chosen with equal probability each, then it should be fine. I asked because I couldn't seem to find any documentation that explained how the available unit would be chosen, so I thought maybe the first unit would always be chosen first by default, which would not be what I want. Thank you. :) On Wed, May 16, 2012 at 9:34 AM, Tony Vignaux &lt;vignaux&lt; at &gt;gmail.com&gt; wrote: </pre> Josephine Lim 2012-05-16T12:20:49 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.comp.python.simpy.user