Performance Scaling of Multi-objective Evolutionary Algorithms

By V. Khare, Xin Yao, Kalyanmoy Deb. International Conference on Evolutionary Multi-Criterion Optimization, 2003

495 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computational Theory and Mathematics

View Paper: https://www.semanticscholar.org/paper/e941de38bcdaa992badfdea7f3377321a2d907cc

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Yaochu Jin
2. Gary G. Yen
3. Ran Cheng
4. Xin Yao
5. Yanan Sun
6. Shengxiang Yang
7. Xing-yi Zhang
8. Ye Tian

In Computational Theory and Mathematics

1. Yaochu Jin
2. Gary G. Yen
3. Ran Cheng
4. Shengxiang Yang
5. Xin Yao
6. Yanan Sun
7. Xing-yi Zhang
8. Ye Tian

Citations in the corpus

• IGD Indicator-Based Evolutionary Algorithm for Many-Objective Optimization Problems (2018, background): “M ANY-OBJECTIVE optimization problems (MaOPs) refer to the optimization tasks involving m (i.e., m > 3) conflicting objectives to be optimized concurrently [1].”
• Multi-objective optimization using evolutionary algorithms (2001, result): “An earlier study [50] with eight objectives revealed somewhat negative results.”
• Performance of Decomposition-Based Many-Objective Algorithms Strongly Depends on Pareto Front Shapes (2017, methodology): “The DTLZ, WFG and knapsack problems were repeatedly used for demonstrating difficulties of many-objective optimization for the Pareto dominance-based EMO algorithms [13]-[16].”
• An Evolutionary Many-Objective Optimization Algorithm Based on Dominance and Decomposition (2015, background): “Confronted by the mentioned difficulties, the performance of most current EMO algorithms, which have already demonstrated their capabilities for two- and three-objective problems, deteriorate significantly when more than three objectives are involved [20].”
• On the Evolutionary Optimization of Many Conflicting Objectives (2007, methodology): “Like the polynomial mutation operator, the SBX recombination operator has also been previously used in the D2 algorithm on DTLZ2 [6], [5].”
• Two_Arch2: An Improved Two-Archive Algorithm for Many-Objective Optimization (2015, background): “IN the real world, many problems [1], [2] can be regarded as many-objective optimization problems (ManyOPs) [3], [4], a category of multiobjective optimization problems (MOPs) [5] with more than three objectives.”
• Multiobjective Immune Algorithm with Nondominated Neighbor-Based Selection (2008, methodology): “As the reference (Khare et al., 2003), in Section 4.2, we apply this metric only to the final Pareto-optimal set obtained by an MOEA to evaluate its performance.”
• Evaluating the -Domination Based Multi-Objective Evolutionary Algorithm for a Quick Computation of Pareto-Optimal Solutions (2005, background): “Although in two-objective optimization problems, the diversity obtained by these two MOEAs was reported to be almost similar (Deb et al., 2002a), a remarkably large difference was evident in solving three or more objective problems (Deb et al., 2002b; Khare et al., 2003) .”
• A New Two-Stage Evolutionary Algorithm for Many-Objective Optimization (2019, unknown): “, [0, 1] and [1, 0].”
• An experimental methodology for response surface optimization methods (2011, unknown): “To date, however, many algorithm evaluations have used only a small number (i.e. fewer than twenty) hand-constructed test functions, although there is work on expanding this number (Khare et al. 2003; Deb et al. 2005).”
• Balancing Convergence and Diversity in Decomposition-Based Many-Objective Optimizers (2016, background): “Unfortunately, experimental and analytical results [8], [9] have indicated that MaOPs would pose serious difficulties to the existing MOEAs, particularly for the popular Pareto-dominance-based MOEAs, e.g., nondominated sorting genetic algorithm II (NSGA-II) [10] and strength Pareto evolutionary…”
• Infill sampling criteria for surrogate-based optimization with constraint handling (2012, background): “NSGA-II also has inherent problems when dealing with more than a few objectives (Khare et al. 2003, Praditwong and Yao 2007, Schütze et al. 2010).”
• A survey on multi-objective evolutionary algorithms for many-objective problems (2014, background): “[57]werethefirsttousetheDTLZsuite [25]tocomparetheperformanceofthreeMOEAsinamany-objectivecontext.”
• A Competitive-Cooperative Coevolutionary Paradigm for Dynamic Multiobjective Optimization (2009, result): “Recent works [26], [ 34 ] showed that the performances of MOEAs often do not scale well with respect to the number of objectives.”
• Evolutionary many-objective optimization: A short review (2008, background): “, the deterioration of the search ability of EMO algorithms by the increase in the number of objectives) has been pointed out in a number of studies (for early studies, see [7], [8]).”
• Shift-Based Density Estimation for Pareto-Based Algorithms in Many-Objective Optimization (2014, methodology): “Using a grid technique to maintain diversity, PESA-II has been found to outperform NSGA-II and SPEA2 in many-objective optimization [46].”