In Progress

Probability of exclusive gateway, simulation bug.

I think there is little buggy, setting and calculation of probability of exclusive gateway simulation (process Validation).

1. Difficult to use sliding button. There is problem to set probability like 98:2, 99:1 It depends how fast setting button is moved. If fast, there is a problem (and maybe which part of setting button is caught with mouse cursor too). Maybe its caused by low machine performance.

2. Is impossible to set by mouse probability 100:0 (in case of temporary testing etc.). It must be done by highlighting the setting button and using arrow keys. And moreover to set 100:0 user must go with arrows to right and it ends on 99:1 then, when left arrow is pressed 100:0 is set. (=little bit of user hell)

3. Calculation. I) If i set for example 1000 going on gateway with setting 99:1 it ends with the same result 985:15 independently how many times i run simulation again. If I'm wrong please correct me but i think it have to be 990:10. If the calculation includes the chance and variance, then it must be changed with each and the results must vary - it doesn't.

II)Second think is when i set 1:99 it ends with cases result 14:986. When count of cases is lowed from 1000 to 100 the problem escalates. Setting 99:1 -> result 98:2. I think is something in the application code, wrongly calculated cases divided by percentages, using inappropriate date type and rounding etc.Tested on simpliest scheme of exclusive gateway.

P.S. Attach file does not support .bpm files.

Best Answer
photo

Dear Radek,

Please refer to the following link to check the information about

simulation in Bizagi

http://help.bizagi.com/processmodeler/en/index.html?simulation_in_bizagi.htm

Regards

photo
0

Dear Radek,

Please refer to the following link to check the information about

simulation in Bizagi

http://help.bizagi.com/processmodeler/en/index.html?simulation_in_bizagi.htm

Regards

photo
0

Hi. I have observed that the problem with bad calculation happened after try setting 100:0. Then something gone bad and the next time running calculations are miscalculated.I am convinced, it's not working properly. I have analysed and simulated that situation many times in different variant. If you don't want to acknowledge... your business. ;-) Best regards.

photo
0

Hi.

First my sorry. You're right, using replication in simulation in the end calculates logic statistically average values. It suppresses point 3 a) & b). (However i think the implementing of setting seed for random generator is at my first look little superfluous , but maybe it has it's own specific sense)

Second. Only point 1) a and 2) still living. :-)

Regards.

photo
0

Dear Radek,

For point 1 and 2 we will check that and pass this observation to

our product team. Your ideas and suggestions are welcomed and are very

important to improve our products.

Regards.