adevs Activity
A simulation library for discrete event and hybrid dynamic systems.
Brought to you by:
smiz
Activity for adevs
-
smiz committed [r410] on Code
Added arguments to sri example for population and random seed
-
Added spline to compiled library to avoid duplicate function definitions
-
Fixed grammar and spelling errors in the API comments.
-
Added fast interpolator to fmi event test case
-
Event location with interpolation reports interpolated solution at the event instead of computing a new solution with the ode solver
-
Added option to use spline to interpolate rather than invoking the ODE solver at each iteration of the fast event finding algorithm
-
Added new bisection event locator that avoids creeping up on an event
-
Fixed error in the trap truncation error estimate
-
Working example of harmonic compensator with Modelica circuit and control and timed or state events for switching
-
New error control for trap uses Euler starting guess. Avoids extra implicit solutions to get error estimate.
-
Reduced number of derivative function calls in trap by reusing calculations across step calls. Fixed jacobian calculation in FMI so that it works column by column and does not (appear) to cause open modelica FMIs to crash.
-
Added option for no jacobian to FMI. Fixed error estimate in trap. First cut at modelica FMI implementation of harmonic compensator.
-
Euler start for trap. Fixed up test cases for fmi and ode.
-
Trap integrator works with KINSOL.
-
Working towards kinsol nonlinear solver for trap method. Compiles. Doesn't run yet
-
Step size adapts to ensure newton convergence in trap method
-
Added test for existance of jacobian to trap
-
Fixed memory leak in the trapezoidal integrator
-
Added variable step size for error control to the trapezoidal integrator
-
Added fixed trapezoidal integrator and support for calculating jacobians from FMI directional derivatives.
-
Initial active list passed to simulator can contain coupled and atomic models
-
Added example of harmonic conpensation using an inverter
-
Documentation update
-
Code cleanup to enable a clean compile with clang++
-
Fixed a misleading comment in the example
-
Simplified the circuit example by removing some extraneous functionality
-
Added some notes to the glife example and changed it to use the integers for time
-
Added a fast constructor that can operate directly on a list of atomic components supplied by the modeler.
-
Update links in web landing page
-
adevs released /adevs-3.3.zip
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Documentation cleanup, code cleanup, project cleanup
-
Corrected error in FMI postStep that set the time incorrectly.
-
Corrected used of Fmi2SetTime outside of continuous block. Working on correct use of state event indicators in Fmi examples as the relation of event number and event of interest are not fixed until compile time.
-
Added RK2 inspired integrator
-
Updates fmi features to build with the new omc compiler. The robotic arm examples are broken now, and I have no idea why.
-
Automated pass/fail check for ode confluent test by adding assertions
-
Added a 2D CA example
-
Added missing template argument to route method for CellSpace
-
Small improvement to ca model. Some code cleanup for des model
-
Fixed a silly error that had comparison with E inside the call to calculate energy rather than outside
-
Fixed DES model to ensure confluent transitions were handled correctly
-
Working DES model.
-
Correct(?) CA model. DES needs work.
-
First cut at a model of grain growth in materials
-
Improved support for super dense time. The simulator now checks for injecting events in the past. To facilitate using super dense time, the simulator calls that change the model state return the time at which that state is valid. Input can be injected at this time up to and including the time of next event. The first field of super dense time is now a template argument. Updated test cases and examples to reflect this; new test cases test the return type for the state changing calls to the sim...
-
Forgot the new test case for the inputListener method!
-
Added a inputListener method to the EventListener, plus test cases.
-
Added an example of an asynchronous automaton
-
Thanks for your interest. At one time, I maintained a git repository for this project, but you are right that it hasn't been updated in a long time. For what it might be worth, this will create a new git repository from the subversion one. git svn clone https://svn.code.sf.net/p/adevs/code/trunk adevs-code From there, you can use git svn fetch to get fetch updates from the svn repository into your git repository. I used to have the synchronization with github working automatically, but that appears...
-
Lee Van Dyke posted a comment on a wiki page
Hi, I've started porting this to Fedora 31 and would like hand-holding as to host on github. Don't know svn and barely know git. Your git page looks inactive as the last update shows 2015? Thanks in advance, Lee
-
Code cleanup. Fixed up some documentation, fixed some compiler warnings, and got rid of code that isn't maintained anymore
-
Hi Michael, bdevs is intended to be a simpler, friendlier version of adevs. It is stable enough to use (I'm using it in a project now) but doesn't yet support continuous systems like adevs does. My plan is to continue development of bdevs and maintain adevs as needed. If you decide to use bdevs, I'd be curious to hear your thoughts on it. Thanks for your interest! Jim On Thu, Jun 6, 2019 at 3:51 PM Michael O'Keefe michaelokeefe@users.sourceforge.net wrote: Hi Jim, Curious if you could speak to bdevs...
-
Michael O'Keefe posted a comment on discussion Help
Hi Jim, Curious if you could speak to bdevs which I also found on SourceForge from you. At this point, would you still recommend sticking with adevs for doing PDEVS simulations? Or is bdevs at a stable point? Thank you and thank you for sharing adevs with the world. Sincerely, Michael
-
Code cleanup
-
Added an example of a time managed qemu simulation with gpu passthrough
-
Code cleanup
-
Updated support for interacting with QEMU via shared memory
-
The code at web.ornl.gov/~nutarojj/adevs/ is a stable snapshot of the development code that is in the source forge repository. The snapshots that get put onto source forge often lag the ones at web.ornl.gov/~nutarojj/adevs/, mostly because I'll forget to put up the second post to source forge. I'd recommend either using the code from the web.ornl.gov site or the code at the head of the repository. Jim On Sat, Feb 17, 2018 at 3:22 PM, bergtwvd bergtwvd@users.sourceforge.net wrote: What is the difference...
-
bergtwvd posted a comment on discussion Help
What is the difference with the ADEVS code on the site web.ornl.gov/~nutarojj/adevs/ ? From there I downloaded adevs-3.3, which appears more recent than this repository.
-
Fixed some foolishness with the misuse of a conditional
-
Fixed problem with write blocking on attempt to send device data to qemu
-
Added option to delay start of the emulator in simulation time
-
Removed broken multi scheduler. Fixed build for java by removing openmp arguments to compiler
-
Changed qemu options to work with 2.11
-
Update for new qemu patches
-
Removed check for self influencing models.
-
Fixed problem in which Mealy model had delta_conf called instead of delta_ext when it received input but was not imminent
-
Added float as valid option for time
-
Added an example of an activity scanning framework that will be described in the upcoming book Model Engineering for Simulation
-
-
Placed call to model_transition after reporting...
-
Removed needless write thread for sending data ...
-
Code cleanup to fix problems with building on W...
-
Removed conservative discrete event simulator (...
-
Removed unncessary threading from qemu_rpc and ...
-
Updates to follow qemu capability in main branch
-
Removed critical section from MultiSchedule and...
-
Implements a cut down version of parallel DEVS ...
-
Added parallel computation of state transition ...
-
Removed old parallel code for conservative simu...
-
First attempt to replace existing parallel code...
1 >