Algorithms: Sumit Mishra's divide‐and‐conquer

@inproceedings{ dncs-cec2016,
    author      = {Sumit Mishra and Sriparna Saha and Samrat Mondal},
    title       = {Divide and Conquer Based Non-dominated Sorting
                   for Parallel Environment},
    booktitle   = {Proceedings of Congress on Evolutionary Computation},
    year        = {2016},
    pages       = {4297-4304},
    publisher   = {IEEE},
    langid      = {english}
}

How to get an instance:

• SumitMishraDivideConquer.getDCNS_SS() -- returns an implementation that uses sequential search while determining ranks. The worst-case running time complexity is O(MN^2).
• SumitMishraDivideConquer.getDCNS_BS() -- returns an implementation that uses binary search while determining ranks. The worst-case complexity is also O(MN^2).

The dichotomy on sequential search vs binary search is mostly the same as in the case of ENS_SS versus ENS_BS. The current implementation is always better, often five to ten times, than the original implementation of Sumit Mishra. Despite that, it is still not very competitive compared with some of the algorithms.