@@anylogic-simulation Great work on Pypeline! Effective and easy to use. Can you give an indication on when the library ALPyne will be ready for Reinforcement Learning purposes? (re: 00:52:50 Q&A - About training RL agents w/Pypeline).
Hello Tyller, I've a question regarding to your presentation. I want to perform a simulation which is fed by machine learning predictions like 00:07:10. I am asking whether you have an example or code about it. I hope you respond. Thanks in advance.
Hello and thanks for the question. The model source files for the model “Simple Hospital (AI Testbed),” which is shown at 00:07:10, are available here: github.com/t-wolfeadam/AnyLogic-Pypeline/tree/master/Examples/StandardModels/Simple%20Hospital%20(AI%20Testbed). Also, the official GitHub page for the Pypeline Project could be helpful for you: git.io/al_py
Hello Tyler. Thanks a lot for the information regarding the connector. Im having some issues in passing a population from python back to Anylogic.(a population of 100 with an int array and a double). I am succesfull in passing it back to anylogic as a string, however I cannot seem to find the correct inputs for the fromPopulationJson() to cast it to a population in my model. Do you have any very small (coding) examples for fromPopulationJson and fromAgentJson which I can use as an example? The existing examples go as far as a HashMap, which is not exactly what I need. Namely the Engine engine, Agent owner inputs confuse me a little..
Hi! And thanks for writing. Completely understandable issue - I've been behind on updating the wiki and having better examples of these functions (which I've also simplified in later versions). If you check the "Advanced functions/usage" page on the Pypeline wiki, I added a _lot_ more description and a new demo at the end. Note that you'll need the latest Pypeline version (also updated that) to run the demo properly. If you have any more Qs, please ask them in the discussions tab on the github page so that I get notified :)
And since I can't post links, you can find the page by googling "pypeline github" - should be the first result (or it's in the description at the end); on it, releases are on the middle right and the wiki/discussion tabs are on top.
Absolutely! Since Pypeline just uses your system's Python installation, any library can be used. For an example using specifically numpy and scipy, see the "Machine Optimizer" example model (accessible in AL from: Help > Example models > How-to models > Using Python in AnyLogic). The imports happen on Main's "On startup" field and the logic in the "solveit" function. If you want to put the logic in a dedicated Python script, just follow the format used in many of the other examples (like the "Simple Hospital" model).
Great explanations. Hope you continue with the good work, Tyler!
It was a very useful video to learn how to interact with python. I appreciate all your works on Pypeline library.
Great! You may also be interested in this Python related blog about AnyLogic API - www.anylogic.com/blog/python-api-for-simulations-in-anylogic-cloud/
@@anylogic-simulation Great work on Pypeline! Effective and easy to use. Can you give an indication on when the library ALPyne will be ready for Reinforcement Learning purposes? (re: 00:52:50 Q&A - About training RL agents w/Pypeline).
Hello Tyller,
I've a question regarding to your presentation. I want to perform a simulation which is fed by machine learning predictions like 00:07:10. I am asking whether you have an example or code about it. I hope you respond. Thanks in advance.
Hello and thanks for the question.
The model source files for the model “Simple Hospital (AI Testbed),” which is shown at 00:07:10, are available here: github.com/t-wolfeadam/AnyLogic-Pypeline/tree/master/Examples/StandardModels/Simple%20Hospital%20(AI%20Testbed).
Also, the official GitHub page for the Pypeline Project could be helpful for you: git.io/al_py
Hello Tyler. Thanks a lot for the information regarding the connector.
Im having some issues in passing a population from python back to Anylogic.(a population of 100 with an int array and a double). I am succesfull in passing it back to anylogic as a string, however I cannot seem to find the correct inputs for the fromPopulationJson() to cast it to a population in my model. Do you have any very small (coding) examples for fromPopulationJson and fromAgentJson which I can use as an example? The existing examples go as far as a HashMap, which is not exactly what I need. Namely the Engine engine, Agent owner inputs confuse me a little..
Hi! And thanks for writing. Completely understandable issue - I've been behind on updating the wiki and having better examples of these functions (which I've also simplified in later versions).
If you check the "Advanced functions/usage" page on the Pypeline wiki, I added a _lot_ more description and a new demo at the end. Note that you'll need the latest Pypeline version (also updated that) to run the demo properly. If you have any more Qs, please ask them in the discussions tab on the github page so that I get notified :)
And since I can't post links, you can find the page by googling "pypeline github" - should be the first result (or it's in the description at the end); on it, releases are on the middle right and the wiki/discussion tabs are on top.
Thank's Tyler, here's the link the pypeline github - github.com/t-wolfeadam/AnyLogic-Pypeline
I do have my own question though. How long does it take to be at you guys' level or even higher?
Can you execute a model that uses pypeline where the python script imports and uses important Python libraries such as numpy, scipy and pandas?
Absolutely! Since Pypeline just uses your system's Python installation, any library can be used.
For an example using specifically numpy and scipy, see the "Machine Optimizer" example model (accessible in AL from: Help > Example models > How-to models > Using Python in AnyLogic). The imports happen on Main's "On startup" field and the logic in the "solveit" function.
If you want to put the logic in a dedicated Python script, just follow the format used in many of the other examples (like the "Simple Hospital" model).