Arena Simulation Example Press Operation

Hello everyone, I will explain the problem number 1 of section 5 of Altiok Melamed’s book. It is a press operation, it is a press department of a car factory. In which we receive plates, which are to press the front part and the part rear. These operations can be performed at any order. First the rear and then the front or upside down. But both have to be done.

00:40 - The plates come to us through a distribution exponential mean 5 minutes. Time use of the front press is given by a uniform distribution between 1 and 5 minutes while the rear press, which takes a little longer, between 2 and 6 minutes. The plate will be attached to the queue of the operation that has less number of plates waiting to be pressed. That is, if the plate arrives and there are more plates in queue of the front press operation, it would join the queue of the rear press process and vice versa. The problem finally indicates that this Press works three shifts 24 hours a day.

01:44 - We have to simulate, we have to create an Arena model and then simulate for one year. The following statistics should be calculated. Average stay time of the plates in the press department. The percentage of use of each operation. The mean queue delay for each of the operations. And the average time spent in the press department of those plates that first join the back press operation and then pass to the front press operation Okay, I’ll go on to explain the model, now I have it created, so it is much faster We start with a CREATE, which I have called Plate Arrival The plates come to us following an exponential on average 5 minutes Next, we would go on to assign using an ASSIGN, we would collect an attribute that I have called ArrTime and with the value tnow that will collect the value of the system time at that moment, that is, the input of the plate From here you have to choose one of the two queues, that of the front press process or the rear press process Both processes are defined as indicated by the problem front press process follows a uniform distribution I have created a resource that I have called FRONT And rear press process, same thing, it’s a SEIZE DELAY RELEASE where follows a uniform distribution 2,6 Once these two processes are created, I can use the DECIDES number of plates in two processes.

04:13 - In this case if the number of plates in front press process is less than or equal to the number of plates in back process, will decide the path of front press process. As it is indifferent, I preferred to start with front process since it is the least time it takes. The issue is that the two processes have to be made, if first performs the front then must pass back process. must pass rear process. How are we going to achieve this? Through two DECIDES First of all we are going to assign creating two variables, one named Front Process and the other Rear Process These two variables, in the case to be first on the way of the front press process the value of the variable Front Process will be 1, while the value of the variable Rear Process will be 0 This way when you get to this DECIDE, going through front press process, DECIDE have two paths, in case the rear press process has already been done, its value is 1, will continue. But, if its value is 0, ie, that the process has not rear press, this would follow the path to the rear process.

05:51 - After this process, the plate would go to DECIDE, this DECIDE, and as the front press process It would already be done, it would go ahead and the press process would be finalized. If the plate came to us this way, it would first press the rear part, as the variable Front Process would have assigned a 0, DECIDE would choose the path of the front press process This would DECIDE, and as the variable Rear Process value is 1 it continue forward Both processes end in a RECORD where the time spent in the press process is collected and the system would end Here is another RECORD just after the front press operation, since the exercise asked us the statistic of the time of stay of those plates that had gone through the rear press process and later by the front press process They had made this path. That is why this time we collect it after this DECIDE. In this way, we are already in conditions to simulate the model for one year and see the results. Here we have the complete model, everything I have explained above. We are going to simulate it for a year. To do this, we first go to the run setup.

08:04 - It will replicate it for 365 days, and 24 hours per day what are the three shifts The base units of time will be minutes We will simulate it without animations and we can run it We are going to see the REPORTS and we are going to answer the statistics that the problem asked us problem asked us We see that the average number of plates that have processed have been approximately 105,000 Here we have the stay time, the system already provides it, which is 12.19 minutes or we can see it within the statistics that we have indicated in the model 12.19 the same Those plates that first processed from rear press we see that the press time is lower, this is because the front press process is faster. They also asked us about the waiting times in the queues and what we observed and it was deduced, that the time in the queue front press is 1.6 minutes while that of the rear process is 4.

6 10:00 - Finally, they asked us for the percentages of use of resources, both front as the rear. The time occupation of the front press process is 58% and the other is 72% And with this we would have the problem solved .