The Evolutionary Algorithm is Memetic, using local search to increase search effectiveness. Currently, the local search is implemented by the JUNG framework, as this framework provides an easy way to implement a force-based algorithm. We note that the conversion to JUNG is inefficient, since the graph is now represented twice. Although we have streamlined this process, we intend to remove the redundancy at a later stage.
CEGA is currently in Beta. Research for appropriate fitness functions for different kind of graphs is still under investigation. However the framework is fully functional, so ready for experimentation of your own visualizations!
The CEGA Project
Evolving graphs since 2010
CEGA is an open-source, highly extendable layout plugin for Cytoscape. The CEGA plugin allows you to specify which features of graph visualization (e.g. the number of edge crossings; the differences in edge lengths) are important for your graphs in Cytoscape.
CEGA is based on an Evolutionary Algorithm (using the JGAP framework) which optimizes a user-defined fitness function. This fitness function incorporates fitness measures that are either predefined or created by the user. The figure below gives a graphical representation of the levels of freedom the CEGA plugin provides. The settings menu in Cytoscape also gives you control over the evolutionary algorithm when it comes to settings as population size, number of generations and mutation rate.
The Evolutionary Algorithm is Memetic, using local search to increase search effectiveness. Currently, the local search is implemented by the JUNG framework, as this framework provides an easy way to implement a force-based algorithm. We note that the conversion to JUNG is inefficient, since the graph is now represented twice. Although we have streamlined this process, we intend to remove the redundancy at a later stage.
CEGA is currently in Beta. Research for appropriate fitness functions for different kind of graphs is still under investigation. However the framework is fully functional, so ready for experimentation of your own visualizations!
The Evolutionary Algorithm is Memetic, using local search to increase search effectiveness. Currently, the local search is implemented by the JUNG framework, as this framework provides an easy way to implement a force-based algorithm. We note that the conversion to JUNG is inefficient, since the graph is now represented twice. Although we have streamlined this process, we intend to remove the redundancy at a later stage.
CEGA is currently in Beta. Research for appropriate fitness functions for different kind of graphs is still under investigation. However the framework is fully functional, so ready for experimentation of your own visualizations!