
208
Complex Networks: An Algorithmic Perspective
results were filtered based on clustering size, density and functional homogeneity to
emphasize properties observed in biological complexes.
In essence, the RNSC algorithm is a cost-based local search algorithm based on
the tabu search heuristic [10]. Starting from a random or user-input clustering, it
first performs a search using a simple integer valued function called the naive cost
function as a preprocessor. It then searches the graph using a real valued cost function
called the scaled cost function. The algorithm operates similar to the Kernighan-Lin
algorithm [18] by iteratively moving one node ...