Thank you very much for the video, it has been very helpful. I would like to know if it is possible to optimize the transfer batch in Anylogic Batch, as I have been trying but not yet succeeded. Thank you very much in advance!
I've tried to find this example model named Oil Terminal , but among the example models , this model doesn't exist. Isn't there any other model to do this experiment instead?
We added the Oil Terminal model in AnyLogic 7.3.5, so probably you need to update your AnyLogic. You can conduct an optimization experiment with any other model that has parameters defined in the top-level agent (Main), but to follow this scenario you have to download AL 8.
Hi Gonzalo! You can find it in the example models and also here in the cloud - cloud.anylogic.com/model/67411f29-6575-4568-83a3-74a22cc3eaad?mode=SETTINGS There is more info about Optimization in the help - anylogic.help/anylogic/experiments/optimization.html
Hi, and thank you for the question. A built-in optimization experiment supports single-objective optimization only. The experiment is powered by the OptQuest engine which supports multi-objective optimization - www.opttek.com/doc/v911engine/OptQuest_Engine_Documentation/help.html#multi-objective-optimization You can access OptQuest API in a custom experiment - anylogic.help/anylogic/experiments/custom-experiment.html#custom-experiment and set multi-objective logic. Just to let you know, we added new genetic optimizer - www.anylogic.com/blog/anylogic-8-8-4-genetic-optimization-and-pedestrian-elevators/ from AnyLogic 8.8.4. Please see the 'Optimization in Custom Experiment' example model (in AnyLogic: Help -> Example Models). It shows the implementation of optimization in Custom Experiment and opened for modifications Java class. Feel free to modify the algorithm for your needs.
The video is really good, but when I try to do the optimization Experiment on this model, after downloading it from the cloud, at the end i get a RunTimeException (Error in the model during iteration 1). The iteration ist sometimes 10, 4,2 and so on. How can I fix this? I did not manage to find a solution online yet. Thank you in advance!
Hello, Thank you for the question. We've investigated the issue and attempted to replicate the error using both the Genetic and OptQuest optimization engines. However, we were unable to reproduce the problem on our end. Please try to follow exactly every step from the video again. If you will continue to experience the error after following the steps from the how-to video, please submit your question and the model to support@anylogic.com so we can reproduce the problem on our end and provide you with the solution. We hope this information is helpful to you.
Thank you. I have a question, what method of optimization use in anylogic? I have use GA, PSO and another method for optimization, but I don't find in anylogic optimization experiment. Thanks
Calculating average queue with statistic element is not correct, because its check queue value periodically and can miss moment when value changed. It's like stroboscopic effect, when you checking queue periodically you have risk find it at same value. Better gather this statistic other way. Make event every time queue value changes, add difference of time to an array element, that correspond to queue value. For example if value is 5, you make something like x[5]=x[5]+time()-lastTime. Probably there more easy way, because when you make your agents wait somewhere, you regularly use Queue block, that have automatically gathered property StatsSize.
Hi! Yes, if a train enters and leaves a queue between the update ticks, then the statistics will not reflect this. Accuracy can be increased if a smaller update rate is defined (e.g. < 1 minute instead of 1 hour). This is a simpler way than manually updating the queue size.
Thank you very much for the video, it has been very helpful. I would like to know if it is possible to optimize the transfer batch in Anylogic Batch, as I have been trying but not yet succeeded. Thank you very much in advance!
I've tried to find this example model named Oil Terminal , but among the example models , this model doesn't exist. Isn't there any other model to do this experiment instead?
The AnyLogic Company Thank you for your kindness. Have a good day!
We added the Oil Terminal model in AnyLogic 7.3.5, so probably you need to update your AnyLogic. You can conduct an optimization experiment with any other model that has parameters defined in the top-level agent (Main), but to follow this scenario you have to download AL 8.
Thanks a lot. Excuse me, where can I find this model?
Hi Gonzalo! You can find it in the example models and also here in the cloud - cloud.anylogic.com/model/67411f29-6575-4568-83a3-74a22cc3eaad?mode=SETTINGS
There is more info about Optimization in the help - anylogic.help/anylogic/experiments/optimization.html
@@anylogic-simulation Thank you so much!
Is there any possibilty to implement multi-objective optimization in AnyLogic?
Hi, and thank you for the question.
A built-in optimization experiment supports single-objective optimization only. The experiment is powered by the OptQuest engine which supports multi-objective optimization - www.opttek.com/doc/v911engine/OptQuest_Engine_Documentation/help.html#multi-objective-optimization
You can access OptQuest API in a custom experiment - anylogic.help/anylogic/experiments/custom-experiment.html#custom-experiment and set multi-objective logic.
Just to let you know, we added new genetic optimizer - www.anylogic.com/blog/anylogic-8-8-4-genetic-optimization-and-pedestrian-elevators/ from AnyLogic 8.8.4. Please see the 'Optimization in Custom Experiment' example model (in AnyLogic: Help -> Example Models). It shows the implementation of optimization in Custom Experiment and opened for modifications Java class. Feel free to modify the algorithm for your needs.
The video is really good, but when I try to do the optimization Experiment on this model, after downloading it from the cloud, at the end i get a RunTimeException (Error in the model during iteration 1). The iteration ist sometimes 10, 4,2 and so on. How can I fix this? I did not manage to find a solution online yet. Thank you in advance!
Hello,
Thank you for the question.
We've investigated the issue and attempted to replicate the error using both the Genetic and OptQuest optimization engines. However, we were unable to reproduce the problem on our end.
Please try to follow exactly every step from the video again. If you will continue to experience the error after following the steps from the how-to video, please submit your question and the model to support@anylogic.com so we can reproduce the problem on our end and provide you with the solution.
We hope this information is helpful to you.
Thank you. I have a question, what method of optimization use in anylogic? I have use GA, PSO and another method for optimization, but I don't find in anylogic optimization experiment. Thanks
We use OptQuest by OptTek Systems, you can find the author's white papers with the method description here: www.opttek.com/products/optquest/
Okey, thanks
Calculating average queue with statistic element is not correct, because its check queue value periodically and can miss moment when value changed. It's like stroboscopic effect, when you checking queue periodically you have risk find it at same value. Better gather this statistic other way. Make event every time queue value changes, add difference of time to an array element, that correspond to queue value. For example if value is 5, you make something like x[5]=x[5]+time()-lastTime. Probably there more easy way, because when you make your agents wait somewhere, you regularly use Queue block, that have automatically gathered property StatsSize.
Hi! Yes, if a train enters and leaves a queue between the update ticks, then the statistics will not reflect this. Accuracy can be increased if a smaller update rate is defined (e.g. < 1 minute instead of 1 hour). This is a simpler way than manually updating the queue size.