Hypervolume is widely used as a performance indicator in the field of evolutionary multi-objective optimization (EMO). It is used not only for performance evaluation of EMO algorithms (EMOAs), but also in indicator-based EMOAs to guide the search. Since its initial proposal in the late 1990s, a wide variety of studies have been done on various topics including hypervolume calculation, optimal μ-distribution, subset selection, hypervolume-based EMOAs, and extensions of the hypervolume indicator. However, currently there is no work to systematically survey the hypervolume indicator for these topics whereas it has been frequently used in the EMO field. This paper aims to fill this gap and provide a comprehensive survey on the hypervolume indicator. We expect that this survey will help EMO researchers to understand the hypervolume indicator more deeply and thoroughly, and promote further utilization of the hypervolume indicator in the EMO field.
7 January 2020 in IEEE Transactions on Evolutionary Computation
In this paper, a new hypervolume-based evolutionary multi-objective optimization algorithm (EMOA), namely R2HCA-EMOA (R2-based Hypervolume Contribution Approximation EMOA), is proposed for many-objective optimization. The core idea of the algorithm is to use an R2 indicator variant to approximate the hypervolume contribution. The basic framework of the proposed algorithm is the same as SMS-EMOA. In order to make the algorithm computationally efficient, a utility tensor structure is introduced for the calculation of the R2 indicator variant. Moreover, a normalization mechanism is incorporated into R2HCA-EMOA to enhance the performance of the algorithm. Through experimental studies, R2HCA-EMOA is compared with three hypervolume-based EMOAs and several other state-of-the-art EMOAs on 5-, 10-and 15-objective DTLZ, WFG problems and their minus versions. Our results show that R2HCA-EMOA is more efficient than the other hypervolume-based EMOAs, and is superior to all the compared state-of-the-art EMOAs.
9 August 2021 in IEEE Transactions on Evolutionary Computation
Subset selection plays an important role in the field of evolutionary multiobjective optimization (EMO). Especially, in an EMO algorithm with an unbounded external archive (UEA), subset selection is an essential post-processing procedure to select a prespecified number of solutions as the final result. In this article, we discuss the efficiency of greedy subset selection for the hypervolume, inverted generational distance (IGD), and IGD plus (IGD+) indicators. Greedy algorithms usually efficiently handle the subset selection. However, when a large number of solutions are given (e.g., subset selection from tens of thousands of solutions in a UEA), they often become time consuming. Our idea is to use the submodular property, which is known for the hypervolume indicator, to improve their efficiency. First, we prove that the IGD and IGD+ indicators are also submodular. Next, based on the submodular property, we propose an efficient greedy inclusion algorithm for each indicator. We demonstrate through computational experiments that the proposed algorithms are much faster than the standard greedy subset selection algorithms. The proposed algorithms also help the research on performance indicators.
In the last two decades, the non-dominated sorting genetic algorithm II (NSGA-II) has been the most widely-used evolutionary multi-objective optimization (EMO) algorithm. However, its performance on a wide variety of many-objective test problems has not been examined in the literature. It has been implicitly assumed by EMO researchers that NSGA-II does not work well on many-objective problems. As a result, NSGA-II has always been excluded from performance comparison with recently proposed many-objective EMO algorithms. Recently, it was pointed out that the performance of NSGA-II on many-objective problems is not always bad. In fact, the poor performance of NSGA-II on many-objective problems is mainly due to the existence of dominance resistant solutions. In this article, we show that the negative effect of the dominance resistant solutions can be remedied by slightly modifying objective values of many-objective problems in NSGA-II. Experimental results show that the modified NSGA-II works well on a wide variety of many-objective test problems.
4 April 2019 in IEEE Transactions on Evolutionary Computation
In this letter, a new hypervolume contribution approximation method is proposed which is formulated as an R2 indicator. The basic idea of the proposed method is to use different line segments only in the hypervolume contribution region for the hypervolume contribution approximation. Comparing with a traditional method which is based on the R2 indicator to approximate the hypervolume, the new method can directly approximate the hypervolume contribution and will utilize all the direction vectors only in the hypervolume contribution region. The new method, the traditional method and the Monte Carlo sampling method together with two exact methods are compared through comprehensive experiments. Our results show the advantages of the new method over the other methods. Comparing with the other two approximation methods, the new method achieves the best performance for comparing hypervolume contributions of different solutions and identifying the solution with the smallest hypervolume contribution. Comparing with the exact methods, the new method is computationally efficient in high-dimensional spaces where the exact methods are impractical to use.
The multiobjective evolutionary algorithm based on decomposition (MOEA/D) with the penalty-based boundary intersection (PBI) function (denoted as MOEA/D-PBI) has been frequently used in many studies in the literature. One essential issue in MOEA/D-PBI is its penalty parameter value specification. However, it is not easy to specify the penalty parameter value appropriately. This is because MOEA/D-PBI shows different search behavior when the penalty parameter values are different. The PBI function with a small penalty parameter value is good for convergence. However, the PBI function with a large value of penalty parameter is needed to preserve the diversity and uniformity of solutions. Although some methods for adapting the penalty parameter value for each weight vector have been proposed, they usually lead to slow convergence. In this article, we propose the idea of using two different values of penalty parameter simultaneously in MOEA/D-PBI. Although the idea is simple, the proposed algorithm is able to utilize both the convergence ability of a small penalty parameter value and the diversification ability of a large penalty parameter value of the PBI function. Experimental results demonstrate that the proposed algorithm works well on a wide range of test problems.