ABSTRACT
Non-dominated sorting is a procedure which is heavily used by many evolutionary multiobjective algorithms. Despite it can be done in polynomial time, it remains an asymptotical bottleneck in many of these algorithms. What is more, some algorithms for non-dominated sorting, which are fast on average, are ridiculously slow in the worst case. Here we prove that a recent algorithm, the dominance degree approach from the paper "Ranking Vectors by Means of the Dominance Degree Matrix" by Zhou et al., has the worst-case time complexity of Θ(M N2 + N3). The Θ(N3) term, which is unusually high for this problem, indicates that this particular algorithm can not be recommended for performance-critical applications.
- Kalyanmoy Deb, Amrit Pratap, Sameer Agarwal, and T. Meyarivan. 2002. A Fast and Elitist Multiobjective Genetic Algorithm: NSGA-II. IEEE Transactions on Evolutionary Computation 6, 2 (2002), 182--197.Google ScholarDigital Library
- Julia Handl, Simon C. Lovell, and Joshua Knowles. 2008. Multiobjectivization by Decomposition of Scalar Cost Functions. In Parallel Problem Solving from Nature - PPSN X. Number 5199 in Lecture Notes in Computer Science. 31--40.Google Scholar
- Andrzej Jaszkiewicz and Thibaut Lust. 2018. ND-Tree-Based Update: A Fast Algorithm for the Dynamic Non-Dominance Problem. IEEE Transactions on Evolutionary Computation 22, 5 (2018), 778--791.Google ScholarCross Ref
- Yuren Zhou, Zefeng Chen, and Jun Zhang. 2017. Ranking Vectors by Means of the Dominance Degree Matrix. IEEE Transactions on Evolutionary Computation 21, 1 (2017), 34--51.Google ScholarDigital Library
Index Terms
- Time complexity analysis of the dominance degree approach for non-dominated sorting
Recommendations
Time complexity analysis of the deductive sort in the best case
GECCO '21: Proceedings of the Genetic and Evolutionary Computation Conference CompanionNon-dominated sorting is one of the important step in multiobjective evolutionary algorithms (MOEAs) which are based on Pareto dominance concept. Though non-dominated sorting can be performed in polynomial time, it remains an asymptotical bottleneck in ...
Approximate non-dominated sorting for evolutionary many-objective optimization
Non-dominated sorting has widely been adopted and shown to be very effective in dominance based evolutionary multi-objective optimization where the number of objectives is two or three. In dealing with many-objective optimization problems, where the ...
Counterexample to the best-case running time of efficient non-dominated sorting algorithm
GECCO '22: Proceedings of the Genetic and Evolutionary Computation Conference CompanionNon-dominated sorting is one of the important steps in Pareto dominance-based multiobjective evolutionary algorithms. In this paper, we show that the number of comparisons in the best case of an approach, Efficient Non-dominated Sort (ENS) from the paper ...
Comments