Collect the total flow time of two itemtype and many replications

Hi ~~

Problem definition:

There are two itemtypes in the model. I'd like to collect the total flow time of each itemtype and write in a global table. And I will design a Experimenter with many scenarios. And there are 30 replications for each scenario. So I want to collect the total flow time of each itemtype for each replication and scenario (like the global_table.xls in the files).

I set two labels of the item.
One is at the Entry of the queue. The other is at the Entry of the sink.


I know just how to calculate the total flow time and write in the global table. But I have two problems.

(1) I don't know how to divide to collect flow time of each itemtype.

(2) Collect the data for each replication and scenario and write into a global table.

Thank a lot.

Hi Vic,

Let the item get type 1, 2, 3, and so.
In the exit I would use this code

int type=getitemtype(item);
double staytime=gettablenum("mytable",type,2)+getlabelnum(item,"end_time")-getlabelnum(item,"start_time");
settablenum("mytable",1,type,staytime);

Do not forgot and change your table to the number of itemtypes you have.

To export the table, you do it this way.
Under Experimenter->Advanced -> End of Replication, add this code:

treenode tabell=node("/Tools/GlobalTables/mytable",model());
string filename="mytable";
filename = concat(filename , "_S" , numtostring(scenario,0,0) , "_R" , numtostring(replication,0,0) , ".csv");
string directory = modeldir();
if(stringlen(directory) < 3)
    directory = documentsdir();
filename=concat(directory,filename);
exporttable(node(">variables/data",tabell),filename,1,1);

Now you will get text files with the total flowtime, there will be one file for each replciation and scenario.
If you have 3 replication and 3 scenario you will get 9 files.

Lars-Olof

Thanks so mach, Lars-Olof.

I want to collect the input of each itemtype. I write the code, but it doesn't work.

Code:
------------------------------------------------------------------
if(type==1)
int input1=getinput(node("/Sink1",model()));
else
int input2=getinput(node("/Sink1",model()));

settablenum("mytable",type,3,input1);
settablenum("mytable",type,3,input2);
-----------------------------------------------------------------
It gives input1 a big wrong number, and gives input2 the number that is the number of input of all item. So It's not what I want >"<.

thx~

Hi Vic,

The getinput will only return the total of what has arrived to the sink.

If you want to record number of input for each type, use this code:

int count=gettablenum("mytable",type,3);
count++;
settablenum("mytable",type,3,count);

Assuming that column 3 is where you keep how many of each type has entered the sink.

Do not forget to reset the table to 0 on each reset.

Lars-Olof

Thanks so mach, Lars-Olof.

I succeed to build the model.

By the way,

I used the Experimenter-->Performace Measures to collect the data of each replication and scenario.

Use "Statistic by individual object" and add the code:

int FlowTime=10;
rewrite -> int stat =FlowTime
else if(stat == FlowTime) value = gettablenum("mytable",1,2);

It's a easy way to catch the data of each replication and scenario.

Lars-Olof, thanks so mach again.

Hi,

isn't there any statistic/reporting function that shows the flowtime/throughput time distribution of all flowitems entered and exited the model (of one scenario replication)?

Like a bar chart that shows the fraction of flowitem with a certain throughput time class (100-120 time units, 120-140 time units, etc.)

Thanks!

Congshi