Simulates consensus dynamics based off of a provided graph-Laplacian matrix. For each simulation random points are generated in (x,y)
coordinates from the range [-50,50] and are updated with each time step as they propogate towards convergence. That convergence is directly impacted by the structure of the corresponding graph that is being modeled.
With each graph the eigenvalues are also listed, where the Fiedler eigenvalue is the first non-zero eigenvalue. You can see the dynamics where with lower Fiedler eigenvalues the graph takes longer to converge.
Use the Select Graph
dropdown to choose which graph to model.
Click the Reset
button to generate new points for the same graph structure.
Feedback can be played live by using the Play
and Pause
buttons. When playing, the points will update for each iteration until they reach relative convergence (where the change is close to zero between iterations).
Iterations can also be forwarded manually by pressing the Next
button.
Note: The x axis and y axis are not currently on the same scale.
forked from lwthatcher's block: Consensus Dynamics Diffusion
Modified http://www.numericjs.com/lib/numeric-1.2.6.js to a secure url
https://d3js.org/d3.v4.min.js
https://www.numericjs.com/lib/numeric-1.2.6.js
https://ajax.googleapis.com/ajax/libs/jquery/3.1.1/jquery.min.js
https://maxcdn.bootstrapcdn.com/bootstrap/3.3.7/js/bootstrap.min.js