<< Click to Display Table of Contents >> ## Level 2 - Throughput time analysis |

The second level of the simulation is useful in measuring end-to-end process time.

Here, resources are not included; Bizagi assumes an infinite capacity to avoid delays in the process flow. This is the best case scenario under the given flow and processing times.

In addition to the information specified in the previous level, the following must be defined in the throughput Time Analysis:

Arrival interval time: Defines the time interval between token generation. Tokens will be created until the max. arrival count is reached. This applies to Start Events, Activities that start processes or Timer Events.

Select the Start Event of the process and click the Gear icon on the pie menu. Set the value for the control.

•One option is to define the arrival interval time as a constant by entering a value. The time units for this value are defined in the property Base Time unit in the scenario´s configuration.

In the following image tokens are generated every 5 minutes.

•Alternatively, define a statistical distribution. Click the advanced icon alongside the field to view and select a distribution.

Once selected, you will be able to set the parameters of the distribution.

In the following image the time between generation of tokens is exponentially distributed with mean equals to 5 minutes. Tokens will be generated until 100 are reached.

Bizagi will warn when the configuration of a Normal distribution might obtain negative numbers, to avoid errors. |

Processing times: Defines the amount of time an Activity or Event needs to process a token. That is, it defines a service time period from the moment a token arrives at an Activity or Event until it is executed.

When the process contains embedded sub-processes, regard the following:

•If the sub-process has activities, assign time to each activity and not to the sub-process to calculate the process time from the execution of each activity.

•If the sub-process does not have activities, assign a estimated process time to it.

Click the Activity or Event. Select the Clock on the pie menu, and enter a processing period in the Time Control.

•You have the option of defining the processing time as a constant by entering values in the corresponding units.

•Alternatively, define a statistical distribution. Click the advanced icon alongside the field to view and select a distribution.

Once selected, you will be able to set the parameters of the distribution.

In the following image the processing time of a token in a specific activity is normally distributed with mean 5 minutes and standard deviation of 3 minutes.

Bizagi will warn when the configuration of a Normal distribution might obtain negative numbers, to avoid errors. |

Once the required data for this level have been defined, click the Run button to execute the simulation.

In the new window, click Start to run the simulation.

When running a simulation, the following analysis data will display.

•Number of instances completed.

•Average time per activity.

•Total processing time per activity

When the simulation is complete, select Results to view the outcome.

For the Time Analysis level, the results of the simulated outcome will contain the following information:

•Name: Identifies the specific BPM shape for which the results are displayed.

•Type: Identifies the element type of the BPM shape.

•Instances completed: Indicates the number of tokens that were processed during the simulation.

•Instances started: Indicates the number of tokens that arrived at the shape.

•Minimum time: Indicates the minimum amount of time that any of the simulated tokens lasted inside the BPM shape before its processing.

•Maximum time: Indicates the maximum amount of time that any of the simulated tokens lasted inside the BPM shape before its processing.

•Average time: Indicates the average amount of time that any of the simulated tokens lasted inside the BPM shape before its processing.

When the process model has Inclusive or Exclusive Gateways with activation probabilities, the average time of the process is computed as a weighted average using this probabilities as weights. |

•Total time: Indicates the amount of time the BPM shape employed to process all the simulated tokens.

You can transfer the results report to Excel by clicking the Export to Excel button.

To understand the Throughput time analysis' results better, consider the next simple process.

The process is composed by a Start event, three simple tasks, and an End event. As you may have noticed, there is also an Exclusive Gateway, meaning that Task 2 and Task 3 are executed regarding the probabilities defined in this gateway. Remember that if you do not define probabilities, sequence flows will be equally distributed.

For this process, activation probabilities are assigned in the Exclusive Gateway as follows:

To simulate the process model, an scenario is created with the following parameters:

•Number of tokens created (Max. arrival count configured in the Start event): 100 tokens.

•Processing time for Task 1: 1 hour.

•Processing time for Task 2: 2 hours.

•Processing time for Task 3: 3 hours.

Additionally, the scenario properties are defined as follows:

Once the simulation scenario is run, the following results are obtained.

Starting with the tasks, you can observe that the Minimum time, Maximum time, and Average time matches the respective task processing time. Moreover, you can see that Total time indicates the amount of time each task employed to process all the simulated tokens that went through it. Specifically, the amount of time employed by Task 1 to process 100 tokens, Task 2 to process 21 tokens, and Task 3 to process 79 tokens.

Calculating manually the task's total time:

•For Task 1 you have 100 tokens to process with a processing time of 1 hour per token. This means that the 100 tokens will be processed in 100 tokens x 1 hour/token = 100 hours. Converting 100 hours to days and hours, you obtain 4 days (4 days x 24 hours/day = 96 hours) and 4 hours (100 hours - 96 hours).

•Similarly, for Task 2 you have 21 tokens to process with a processing time of 2 hours per token. The 21 tokens will be processed in 21 tokens x 2 hour/token = 42 hours. Converting 42 hours to days and hours, you obtain 1 day (1 day x 24 hours/day = 24 hours) and 18 hours (42 hours - 24 hours).

•Finally, for Task 3 you have 79 tokens to process with a processing time of 3 hours per token. 79 tokens will be processed by this task in 79 tokens x 3 hour/token = 237 hours. Converting 237 hours to days and hours, you obtain 9 days (9 days x 24 hours/day = 216 hours) and 21 hours (237 hours - 216 hours).

Now, consider the results obtained for the whole process (Process 1). The first thing you may have noticed is that Minimum time, Maximum time, and Average time values are different form each other. Regarding their definitions:

•Minimum time indicates the minimum amount of time that any of the simulated tokens lasted inside the process. In this case, this time is 3 hours and corresponds to a token that is processed first by Task 1 (1 hour), and then by Task 2 (2 hours).

•Maximum time indicates the maximum amount of time that any of the simulated tokens lasted inside the process. For this process model, this time is 4 hours and its consistent with the time a token lasts in Task 1 (1 hour), and Task 3 (3 hours).

•Average time indicates the average amount of time that any of the simulated tokens lasted inside the process. Bearing in mind that Process 1 has an Exclusive Gateway with activation probabilities, the average time must be calculated as a weighted average. Considering token distribution among the three tasks, the average time is computed as follows: (1 hour x 100 tokens + 2 hours x 21 tokens + 3 hours x 79 tokens)/100 tokens = 3,79 hours. Converting 3,79 hours to hours, minutes and seconds, you obtain 3 hours, 47 minutes (0,79 hours x 60 minutes/hour = 47,4 minutes) and 24 seconds (0,4 minutes x 60 seconds/minute = 24 seconds).

Lastly, the Total time of the process corresponds to the sum of the tasks' total time, that is, 100 hours + 42 hours + 237 hours = 379 hours. Converting 379 hours to days and hours, you obtain 15 days (15 days x 24 hours/day = 360 hours) and 19 hours (379 hours - 360 hours).

To provide a general insight into processing times, the emergency response department has decided to perform a time analysis.

For this analysis the following assumptions have been made:

•Necessary resources to perform activities have infinite capacity.

•The expected time between reports is 5 minutes.

•The simulation will evaluate a period of 1 week.

•The estimated processing times for each of the activities are fixed as shown in the next table:

Activity |
Processing time (min) |
---|---|

Receive emergency report |
4 |

Classify Triage |
5 |

Manage patient entry |
11 |

Pick up patient |
20 |

Arrive at patient place QAV |
7 |

Arrive at patient place BA |
10 |

Authorize entry |
4 |

1. Define trigger times. To do so, click the Start Event and then the Gear icon on the pie menu.

For this example, the expected time between reports is 5 minutes, so set the time to this value. Note the value entered is in minutes.

For more information about units please refer to Scenarios.

2. Define the Activities' processing times.

Click the Activity, select Clock on the pie menu and set a value for the Time control.

In the following image the processing time for the first activity is set to 4 minutes.

3. Once all the processing times have been defined, run the simulation. Click the Run button.

Note the simulation shows analysis findings for each shape in real time as it executes, such as average time, total processing time and the number of completed instances.

4. When the simulation is finished, select Results to view the outcome.

As we mentioned before, the results on this level give us a general insight into the cycle time of the process. For this specific case we are able to identify the expected time a person has to wait from the moment the emergency is reported until medical attention is received.

Based on the results, we can conclude:

•A patient waits at least 16 minutes before receiving medical attention.

•A patient waits no more than 33 minutes for medical attention.

•The expected time a patient waits to receive medical attention is 25 minutes and 3 seconds.