I thought it would be something easy to find, but I've been struggling to find an answer on this on google for 40 minutes...
Hey! The tutorial on Boundary Value Problems states that: "The third argument of BVProblem is the initial guess of the solution, which is constant in this example." However, it does not say how to give a non constant initial solution to the solver. I mean something like, instead of u(t)=pi/2 and u'(t)=pi/2 for all t, you have something like u(t)=sin(t) and u'(t)=cos(t). Thanks!
Hello DiffEq community.
I plan to apply for GSoC 2022 and I would like to contribute to SciML/OrdinaryDiffEq.jl#1063
I'd be thankful if you could help me with my application and help me make relevant decisions during this project.
Rahul
I plan to apply for GSoC 2022 and I would like to contribute to SciML/OrdinaryDiffEq.jl#1063
I'd be thankful if you could help me with my application and help me make relevant decisions during this project.
Rahul
@dynamic-queries as a starer PR, you should try doing SciML/OrdinaryDiffEq.jl#1601 and write about that in your application. http://devdocs.sciml.ai/latest/contributing/adding_algorithms/ would be helpful.
7 replies
Hello everyone! SciMLCon is starting soon! Information below!
Hi SciMLCon attendees,
This is to quickly recap the gameplan for today at SciMLCon 22! The conference will take place on YouTube + Discord/Slack from 9 AM to 6 PM EDT on March 23rd 2022. If you have not registered yet, please sign up for the free ticket here at EventBrite.
All talks will be livestreamed on YouTube, on the Julia programming language channel at the following link https://www.youtube.com/watch?v=NSIAfccnq-0. Attendees are encouraged to join the conference chat platform Discord to interact with other attendees and ask questions after each talk.
To join the server, click on the following URL: https://discord.gg/PRqc3NehEr. Make sure to find the #scimlcon channel. We have also created #scimlcon on the Julia Slack where you can find attendees and ask questions too.
Speakers, check your emails for the calendar invite for when you should join the Streamyard platform. If there are any questions on anything else, join the chats and ask away!
Have a great conference!
Best,
The SciMLCon 22 Organizing Committee
I try to use
diffeqpy.WienerProcess(1,1) but it seems that this does not work
7 replies
Hi, first, congrats for the SciMLCon for people who participated/organized, I had lots of fun following some of the talks.
I am trying to implement a callback function with StepsizeLimiter and I am a bit confused about the function dtFE I have to give. Should this function return my max timestep at the end of it?
I am trying to implement a callback function with StepsizeLimiter and I am a bit confused about the function dtFE I have to give. Should this function return my max timestep at the end of it?
2 replies
Hey everyone! I was looking forward to participating in GSoC this year, and one particular project that really caught my eye was - Physics-Informed Neural Networks (PINNs) and Solving Differential Equations with Deep Learning.
Hopefully, this is the right place to discuss this. If not please guide me to a suitable channel and apologies for any disturbance caused.
A little background about me - I have a deep interest in mathematical subjects be it Calculus, Statistics, Linear Algebra, or anything else. Python has been my primary language for quite some time now, however as I got into deep learning, I have been exploring Julia too lately. Any input on how to move forward would be really appreciated. I have started reading about PINNs and relevant papers and found them quite interesting.
Hopefully, this is the right place to discuss this. If not please guide me to a suitable channel and apologies for any disturbance caused.
A little background about me - I have a deep interest in mathematical subjects be it Calculus, Statistics, Linear Algebra, or anything else. Python has been my primary language for quite some time now, however as I got into deep learning, I have been exploring Julia too lately. Any input on how to move forward would be really appreciated. I have started reading about PINNs and relevant papers and found them quite interesting.
3 replies
Hi everyone, I have one question about the tolerance (Rel_tol and abs_tol) setting when using the direct-dense matrix solvers. I have met one very strange issue when testing different tolerance settings. My code can work normally if I used Rel_tol=1e-8 and Abs_tol=1e-10, even though the code is slow. Then when I tried to use small precision, like rel_tol=1e-6 and abs_tol=1e-8, then the code would not provide any outputs after a long time, there are no errors even. I have no idea why this issue occurs. Would you like to give me some hints and suggestions? Thank you!
@GitterIRCbot Hello, I have one question about the tolerance (Rel_tol and abs_tol) setting when using the direct-dense matrix solvers. I have met one very strange issue when testing different tolerance settings. My code can work normally if I used Rel_tol=1e-8 and Abs_tol=1e-10, even though the code is slow. Then when I tried to use small precision, like rel_tol=1e-6 and abs_tol=1e-8, then the code would not provide any outputs after a long time, there are no errors even. I have no idea why this issue occurs. Would you like to give me some hints and suggestions? Thank you!
the error isERROR: LoadError: Unsatisfiable requirements detected for package OrdinaryDiffEq [1dea7af3]:
OrdinaryDiffEq [1dea7af3] log:
├─possible versions are: [0.0.1-0.0.5, 0.1.0-0.1.1, 0.2.0-0.2.1, 0.3.0-0.3.1, 0.4.0-0.4.2, 0.5.0, 0.6.0, 1.0.0-1.0.2, 1.1.0, 1.2.0, 1.3.0-1.3.1, 1.4.0-1.4.1, 1.5.0, 1.6.0-1.6.1, 1.7.0, 1.8.0, 2.0.0-2.0.3, 2.1.0, 2.2.0-2.2.1, 2.3.1-2.3.2, 2.4.0, 2.5.0, 2.6.0-2.6.1, 2.7.0-2.7.1, 2.8.0, 2.9.0, 2.10.0, 2.11.0-2.11.3, 2.12.0, 2.13.0, 2.14.0, 2.15.0, 2.16.0, 2.17.0, 2.18.0, 2.19.0-2.19.1,
2.20.0, 2.21.0-2.21.2, 2.22.0, 2.23.0, 2.24.0-2.24.1, 2.25.0-2.25.2, 2.26.0, 2.27.0, 2.28.0, 2.29.0-2.29.1, 2.30.0, 2.31.0-2.31.1, 2.32.0-2.32.1, 2.33.0-2.33.2, 2.34.0, 2.35.0-2.35.1, 2.36.0, 2.37.0, 3.0.0-3.0.3, 3.1.0-3.1.1, 3.2.0, 3.3.0, 3.4.0, 3.5.0, 3.6.0, 3.7.0, 3.8.0, 3.9.0-3.9.1, 3.10.0, 3.11.0, 3.12.0, 3.13.0, 3.14.0, 3.15.0-3.15.1, 3.16.1, 3.17.0, 3.18.0, 3.19.0-3.19.1, 3.20.0, 3.21.0, 4.0.0, 4.1.0, 4.2.0, 4.3.0, 4.4.0-4.4.1, 4.5.0, 4.6.0, 4.7.0-4.7.1, 4.8.0-4.8.1, 4.9.0, 4.10.0, 4.11.0-4.11.1, 4.12.0-4.12.2] or uninstalled
└─restricted to versions 6.0.0-6 by DifferentialEquations [0c46a032] — no versions left
└─DifferentialEquations [0c46a032] log:
├─possible versions are: 7.1.0 or uninstalled
└─DifferentialEquations [0c46a032] is fixed to version 7.1.0
OrdinaryDiffEq [1dea7af3] log:
├─possible versions are: [0.0.1-0.0.5, 0.1.0-0.1.1, 0.2.0-0.2.1, 0.3.0-0.3.1, 0.4.0-0.4.2, 0.5.0, 0.6.0, 1.0.0-1.0.2, 1.1.0, 1.2.0, 1.3.0-1.3.1, 1.4.0-1.4.1, 1.5.0, 1.6.0-1.6.1, 1.7.0, 1.8.0, 2.0.0-2.0.3, 2.1.0, 2.2.0-2.2.1, 2.3.1-2.3.2, 2.4.0, 2.5.0, 2.6.0-2.6.1, 2.7.0-2.7.1, 2.8.0, 2.9.0, 2.10.0, 2.11.0-2.11.3, 2.12.0, 2.13.0, 2.14.0, 2.15.0, 2.16.0, 2.17.0, 2.18.0, 2.19.0-2.19.1,
2.20.0, 2.21.0-2.21.2, 2.22.0, 2.23.0, 2.24.0-2.24.1, 2.25.0-2.25.2, 2.26.0, 2.27.0, 2.28.0, 2.29.0-2.29.1, 2.30.0, 2.31.0-2.31.1, 2.32.0-2.32.1, 2.33.0-2.33.2, 2.34.0, 2.35.0-2.35.1, 2.36.0, 2.37.0, 3.0.0-3.0.3, 3.1.0-3.1.1, 3.2.0, 3.3.0, 3.4.0, 3.5.0, 3.6.0, 3.7.0, 3.8.0, 3.9.0-3.9.1, 3.10.0, 3.11.0, 3.12.0, 3.13.0, 3.14.0, 3.15.0-3.15.1, 3.16.1, 3.17.0, 3.18.0, 3.19.0-3.19.1, 3.20.0, 3.21.0, 4.0.0, 4.1.0, 4.2.0, 4.3.0, 4.4.0-4.4.1, 4.5.0, 4.6.0, 4.7.0-4.7.1, 4.8.0-4.8.1, 4.9.0, 4.10.0, 4.11.0-4.11.1, 4.12.0-4.12.2] or uninstalled
└─restricted to versions 6.0.0-6 by DifferentialEquations [0c46a032] — no versions left
└─DifferentialEquations [0c46a032] log:
├─possible versions are: 7.1.0 or uninstalled
└─DifferentialEquations [0c46a032] is fixed to version 7.1.0
@ChrisRackauckas Hello Chris, I have one question want to ask you about the CVode solver. In my current code, I have used Cvode(cvode_mem, tout, y0, &t, CV_NORMAL) to solver my ODE function, f, but it seems like the solver has been trapped in a cycle. It always showed go inside the ODE function, f, but do not get the solution. Does it mean the solver is in the internal step? Meanwhile, I tried to artificially terminate the cycle by using CVodeSetMaxNumSteps(cvode_mem, mxsteps=1), but it seems doesn't work, the solver still go inside the ODE function, f, multiple times. So now I am a little bit confusing, if the solver is trapped in the internal step, why after I define the max step =1, it still go inside the ODE function multiple times? Thank you for your answers.
Hello, I have one question want to ask you about the CVode solver. In my current code, I have used Cvode(cvode_mem, tout, y0, &t, CV_NORMAL) to solve my ODE function, f, but it seems like the solver has been trapped in a cycle. It always goes inside the ODE function, f, but does not get the solution. Does it mean the solver is trapped in the internal step? In this case, I tried to artificially terminate the cycle by using CVodeSetMaxNumSteps(cvode_mem, mxsteps=1), but it still doesn't work, the solver still goes inside the ODE function, f, multiple times. So now I am a little bit confusing, if the solver is trapped in the internal step, why after I define the max step =1, does it still go inside the ODE function multiple times? Thank you for your answers.
Hello everyone, I hope this is the right place to ask this question: I want more information on
initialize and finalize options in ContinuousCallback in event handling section. How can I implement them. I did not understand from the official site. My problem is I have some particles that hit a screen at an unknown time and position. I want to save the position and time of hitting. I construct the callback and everything is very good. But I thought maybe from performance point of view, it is more appropriate that turn on the call back at later times not at t=0. because there is much time for the particle to reach the screen. I'll be thankful if you can help me with that.
@GitterIRCbot Hello, I have one question about the CVode solver. In my current code, I have used Cvode(cvode_mem, tout, y0, &t, CV_NORMAL) to solve my ODE function, f, but it seems like the solver has been trapped in a cycle. It always calls the ODE function, f, (I defined one "hello" output in function f if the function is called), but the solver does not get the solution. Does it mean the solver is trapped in the internal step? What is the solver doing in the cycle, does it updates the step size?
To check whether the solver is trapped in the internal step, I tried to artificially terminate the cycle by using CVodeSetMaxNumSteps(cvode_mem, mxsteps=3), but it still doesn't work, the solver still calls the ODE function, f, multiple times. So now I am a little bit confusing, if the solver is trapped in the internal step, why after I define the max step =3, does it still go inside the ODE function multiple times? Thank you for your answers.
[slack] <isaacsas> Maybe https://discourse.julialang.org/t/blog-post-about-my-experiences-with-julia/79976/28 issues about docstrings could get some headway by just adding short docstrings that link to the relevant DifferentialEquations.jl doc section? As @chrisrackauckas mentions there the issue is one probably doesn't want detailed docstrings for every solution type, but we could add short docstrings to each that simply link to the relevant reference section in the main docs. That would at least point users on where to go when looking via the REPL or VS Code.
[slack] <chrisrackauckas> I think just going and adding a link to the solution handling docs for each of them would make sense. And the problem pages can be converted into docstrings.
[slack] <krishnab> Hey folks. Say, I needed to implement a DifferentialEquations callback to handle cases where the ODE and DDE models go outside of certain hard limits. Then I need to estimate some parameters using DiffEqFlux. I was just wondering if having these callbacks will interfere with the optimization routines in DiffEqFlux? I was watching @chrisrackauckas Juliacon 2020 talk and I believe he mentioned something about issues with callback and adjoint calculations, but then again it might fine now? Just wanted to check before I go down that road and run into a dead end.
[slack] <chrisrackauckas> it's fine by now
[slack] <krishnab> Ahh excellent. then I will give it a whirl.
[slack] <chrisrackauckas> it makes it differentiate with respect to time.
[slack] <chrisrackauckas> that should all be done automatically
[slack] <chrisrackauckas> it's like, 2017 old 😅
[slack] <chrisrackauckas> it desperately needs some love.
[slack] <Vedant> yeah i have some PDE bvp work to do. let's talk about it in p2d tomorrow
[slack] <George Gkountouras> Apparently Pro/Max have 2 and Ultra has 4. This is reflected in https://twitter.com/danieldekok/status/1511348597215961093?s=21&t=2v6pZmDWSYGtShH61THtPA.
Gitter, matrix, slack, and zulip are all linked together. Not sure which is the official part.
Hello, this is rather a julia question rather than diffeq. I have a loss function which has one argument, p, to be called by sciml_train, but in some cases I want to add another argument to the loss function, I know I have seen this capability in some video before but couldn’t find it in the past 24 hours. I do appreciate the help.
I'm testing out MTK. Trying to follow this tutorial https://mtk.sciml.ai/stable/tutorials/spring_mass/. When trying to run it i get "ERROR: MethodError: no method matching AbstractFloat(::Type{Term{Float64, Nothing}})". Any idea what might be wrong? (copy/pasted code from the tut)
Hi all, I'm using Catalyst.jl to programatically generate reaction networks which I'm then attempting to solve with DifferentialEquations.jl. I'm attempting to use some of the stiff ODE solvers to overcome some timescale separation issues, but keep getting the following error upon calling solve():
julia: /buildworker/worker/package_linux64/build/src/ccall.cpp:875: jl_cgval_t emit_llvmcall(jl_codectx_t&, jl_value_t**, size_t): Assertion `f->getReturnType() == rettype' failed.
A long traceback follows. This ONLY occurs for stiff ODE solvers (I've tried a range of them, all with the same result), while all the non-stiff solvers I've tried seem to work fine (they become unstable and intentionally stop, which is expected behaviour for the timescale-separated problems I'm giving them). Has anyone experienced this behaviour before? (Catalyst v10.8.0, DifferentialEquations v7.1.0, also separately tried using OrdinaryDiffEq 6.10.0 with same results)