This page was automatically generated by NetLogo 3.1.4.
created with NetLogo
view/download model file: fuzzylumps06.nlogo
fuzzylumps05.nlogo, 06 august 2006
by ralph abraham, UC santa cruz, abraham@vismath.org
with sisir roy, ISI calcutta, sisir@isical.ac.in
In this model, we simulate the dynamical cellular network (DCN) of nodes
and bonds, as described in Requardt and Roy, 2001. Fuzzy lumps (emergent supernodes and superbonds) and the clique graphs will come later.
An initial random distribution of node and bond states evolves according to a dynamical rule. See the model description document (PDF) on our website.
NOTE: The origin (0, 0) for indices (i, k) is at the lower right corner of the graphics window. In this preliminary version there are 18 nodes, numbered from 0 to 17.
There are four display modes:
"bonds" displays bond-states, J_ik, as colored squares above the diagonal
"nodes" displays node-differences, s_ik as colored squares above the diagonal
"digraph" displays a green square (for "1") above and below the diag
"graph" displays a cyan square (for "1") above and below the diag
In each case, the node-states, s_i, are displayed on the diagonal as colored circles.
Just press "setup" and "step" or "go" as usual.
To display attributes:
1. Press go to stop the simulation;
2. Change the display option with the "display-matrix" drop-down menu; and
3. Press the button "update-display" to change the display type.
Again press "step" or "go".
To display the permutation of the current graph:
1. Press "go" to stop the simulation;
2. Press "show-weight-list".
Then three consecutive lines are shown in the "Command Center":
A. the list of pairs, (node-number, node-weight), in order of node-number,
B. the list A sorted by decreasing weights, and
C. the list of node numbers of list B, that is, the permutation of the graph
See if any stable attractors emerge.
Move sliders and begin a fresh run.
Fuzzy lumps are coming later. Our plan is to run the model until an attractor is reached, then export a permutation to an external program such as Combinatorica to find its cliques and weights, and then submit this data to another NetLogo model to self-organize the cliques and weights as supernodes aproximating an isometric embedding.
Agents are useful for the emergence of supernodes
See Nowotny, dissertation.
Requardt and Roy, 2001