Compute the crowding distance of a set of points.

The crowding distance is a measure of spread of solutions in the approximation of the Pareto front. It is used, e.g., in the NSGA-II algorithm as a second selection criterion.

computeCrowdingDistance(x)

Arguments

x	[matrix] Numeric matrix with each column representing a point.

Value

[numeric] Vector of crowding distance values.

References

K. Deb, A. Pratap, S. Agarwal, T. Meyarivan, A fast and elitist multiobjective genetic algorithm: NSGA-II, IEEE Transactions on Evolutionary Computation In Evolutionary Computation, IEEE Transactions on, Vol. 6, No. 2. (07 April 2002), pp. 182-197, doi:10.1109/4235.996017