maximum agents staying in one waiting area & Poisson Distribution

All topics related to questions in Pedestrian Dynamics
Mona_Hung
Posts: 10
Joined: Wednesday 17 October, 2018 - 11:25

Re: maximum agents staying in one waiting area & Poisson Distribution

Post by Mona_Hung »

Dear Ignasi Echaniz,

Thanks a lot for your help. I guess I am able to due with the Poisson distribution and waiting time.
Would I ask what is the difference if I choose repeat time or no repeat like the following picture?
I am not sure how will it affect the model.
figure 1
figure 1
Repeat and no repeat time.jpg (154.3 KiB) Viewed 3290 times
Moreover, about the output evaluation, from the figure 2, we may see that only one experiment would be analysed at one time. Is it possible that the overall results of all experiments I made would be evaluated together? For example, I would like to know the average waiting time of one agent of 100 experiments but not just one.
figure 2
figure 2
result analysis.jpg (117.68 KiB) Viewed 3290 times
Thanks for your kindly help.

Yours gratefully,
Mona
Mona_Hung
Posts: 10
Joined: Wednesday 17 October, 2018 - 11:25

Re: maximum agents staying in one waiting area & Poisson Distribution

Post by Mona_Hung »

And here is my model. Hope you can take a look for me.
In fact, I would like to test the best number of chairs (which is presented by the waiting area in the model) for this area.
For the experiment, I would like to know the average percentage that the chairs are used.
Also, I would like to try to change the parameter of the number of agents and take a look of what will happen. For example, I would like to know how many chairs are needed if the number of agents increase greatly.
Besides, I hope to change the number of chairs in the experiment and try to calculate the best number of chairs provided by the owner.

I heard that some simulation software may do the experiment with variable parameter easily. So, I just wonder whether Pedestrian Dynamics is able to do it or I need to do the experiment and analysis it one by one.

Again, thanks a lot for your help. I do appreciate your effort, thanks.
Attachments
5min_Avenue 0001.mod.bak.bak.bak.bak.bak.bak.bak.bak.bak.bak.mod
(543.1 KiB) Downloaded 105 times
Mona_Hung
Posts: 10
Joined: Wednesday 17 October, 2018 - 11:25

Re: maximum agents staying in one waiting area & Poisson Distribution

Post by Mona_Hung »

Furthermore, would I set different maximum capacity to different waiting areas in the attached file?
The maximum capacity of different areas is as follow:
circle - 7 people
square - 4 people
parallelogram - 16 people

Thanks a lot.
Attachments
K11_001.mod.bak.bak.bak.bak.bak.mod
(384.56 KiB) Downloaded 110 times
User avatar
Ignasi Echaniz
Posts: 18
Joined: Monday 28 May, 2018 - 10:44

Re: maximum agents staying in one waiting area & Poisson Distribution

Post by Ignasi Echaniz »

Hi Mona Hung,
I will answer all your question in chronological order:

1. The difference between RepeatTime and NoRepeat is that you generator will repeat (or not) according to the specified “Repeat Time”. That means that your arrival list (other tab) will be repeated over time or not. This can help you if you want to repeat the same distribution of agents over time. More explanations can be found if you click the question mark in the bottom left corner of the GUI.

2. That’s right. When you load an experiment only refers to that specific experiment. We have some functions to evaluate multiple runs statistics. Might I ask you which Pedestrian Dynamics version are you using?

3. You can use a user table to easily set the capacity of each type of your waiting area (circle, square….). First you will need to use labels to identify which kind of activity is each. Then in the indicative corridor stop condition you will need to set the threshold according to the activity you are checking, and refer this threshold to the user table.
Another option is to not use the indicative corridor and use server queues. There you can set the capacity of each server (activity) . Then you will need to use and modify the activity location assignment to your needs.
I recommend you to check tutorial example models and explanations of activity locations, indicative corridors and basic 4dScript for user tables (Help tab).

4. Experiment parameters can be easily changed every run. To do so, I recommend to use the start code of each run (experiment wizard GUI – see figure-). There, you can easily make a code to change the user table that I just discuss in the previous point. Then each run you will have different capacity of your activities automatically. To make this short code 4ds script I recommend you to take a look in the help of the function “cell”.

5. If you are interested in the pure processes and not in crowd behaviour, I would also recommend to download our software “Enterprise Dynamics”. This software is able to model any virtual process and will be easier (less coding) to get the results you are mainly asking: occupancy of a certain process. There you can also find an example model of how to model a restaurant to determine the optimal number of table according to the demand.

Best regards,

Ignasi Echaniz
Attachments
community.png
community.png (17.94 KiB) Viewed 3281 times
Mona_Hung
Posts: 10
Joined: Wednesday 17 October, 2018 - 11:25

Re: maximum agents staying in one waiting area & Poisson Distribution

Post by Mona_Hung »

Dear Ignasi Echaniz,

Actually I'm still using the trial. Is that the trial do not have the functions to evaluate multiple runs statistics? Thanks.

Yours gratefully,
Mona
User avatar
Ignasi Echaniz
Posts: 18
Joined: Monday 28 May, 2018 - 10:44

Re: maximum agents staying in one waiting area & Poisson Distribution

Post by Ignasi Echaniz »

Hi Mona_Hung,

The version you have allows to statistically analyse multiple runs. To do so, you will need to use the interact (SHIFT + F6). Then use the following code:

AgentStatistics_CalculateOverallOutputStatistics(e1,e2,e3,{e4},{e5}), where
  • e1 = Directory where experiments are saved
  • e2 = Statistics to obtain (STATISTIC_CONFIDENCEINTERVAL, STATISTIC_PERCENTILE, STATISTIC_AVERAGE, STATISTIC_MAX, STATISTIC_MIN)
  • e3 =Inteval of confidence or percentile (if not, 0).
  • e4 = Optional. Number of first replication.
  • e5 = Optional. Last replication.
One Example could be: AgentStatistics_CalculateOverallOutputStatistics([D:\My Documents\Pedestrian Dynamics\3.0 Official Release\PD_Results\Experiment1\],STATISTIC_PERCENTILE, 95)

Once this code is executed, a new windows will appear with the statistics of multiple runs. We are currently working in improving multiple runs features in near future releases.

Kind regards,

Ignasi Echaniz
Post Reply