An enhanced-indicator based many-objective evolutionary algorithm with adaptive reference point
Introduction
Many real-world problems involve more than one conflicting objectives to be optimized simultaneously. These problems are referred to as multi-objective optimization problems (MOPs) [1,2], which can be stated as follows:where is the decision space with being the decision vector, consists of M objectives and Y is the objective space.
In the past three decades, a lot of multi-objective evolutionary algorithms (MOEAs) have been proposed to solve MOPs. These MOEAs have shown promising performance in solving MOPs [3]. However, some recent studies have demonstrated that conventional MOEAs faced with difficulties when handling MOPs with more than three objectives, and these MOPs are further differentiated as many-objective optimization problems (MaOPs) [4,5]. Due to the increasing number of objectives, traditional MOEAs lack selection pressure when solving MaOPs. To enhance the performance of MOEAs in solving MaOPs, a variety of many-objective evolutionary algorithms (MaOEAs) have been proposed. In dominance relation based methods, the conventional Pareto-based methods face difficulties to distinguish solutionsin tackling MaOPs. The most straightforward method is to relax the domination relation to enhance the selection pressure towards the Pareto front (PF) [[6], [7], [8]]. There are also other novel strategies to enhance convergence. Recently, Zhang et al. proposes a knee point driven evolutionary algorithm (KnEA) [9], which introduces a knee point based selection criterion to enhance convergence.
Unlike dominance based MaOEAs, decomposition based MaOEAs provide a promising approach to solve MaOPs [10]. The multiobjective evolutionary algorithm based on decomposition (MOEA/D) [11] is a representative of this class of algorithms. Recently, R. Cheng et al. propose a reference vector guided evolutionary algorithm (RVEA) for MaOPs [12], which decomposes MaOPs into a set of single-objective optimization problems (SOPs) via a set of uniformly distributed reference vectors, such that the candidate solutions can efficiently converge to the optimum of each SOP without considering the conflicts between different objectives. An evolutionary many-objective optimization algorithm using reference point based nondominated sorting approach (NSGA-III) is proposed in Refs. [13], which adopts the reference points based secondary criterion to manage diversity. In Ref. [14], a dynamical decomposition based evolutionary algorithm (DDEA) is proposed to solve MaOPs, which introduces a dynamical decomposition strategy to partition the objective space by solutions themselves without predefined reference vectors.
Indicator-based MaOEAs provide another potential approach to solve MaOPs. Their core idea is to adopt performance indicators of solutions quality measurement as selection criteria in environmental selection [15]. R2 indicator based many-objective metaheuristic-II (MOMBI-II) is proposed in Ref. [16], which proposes a reference point update method to solve the problem that the diversity of the approximation sets is sensitive to the choice of the reference points during normalization. In Ref. [17], an IGD-NS (enhanced inverted generational distance) indicator based evolutionary algorithm, termed AR-MOEA, which proposes a reference point adaptation to handle different types of Pareto fronts. An IGD indicator-based evolutionary algorithm (MaOEA/IGD) is proposed by Yanan Sun et al. [18], which introduces the IGD indicator as selection criteria to select solutions in each generation, and one rank assignment mechanism is proposed to compare the dominance relation of the solutions with the reference points.
These MaOEAs are proposed to solve MaOPs, and the most existing MaOEAs have shown potential performance in solving different types of MaOPs. However, some recent studies point out that the most existing MaOEAs always encounter difficulties in diversity management for many-objective optimization [19,20]. Hence, this paper proposes an enhanced-indicator based many-objective evolutionary algorithm with adaptive reference point, namely EIEA. The major contributions of this paper are summarized as follows:
- 1)
The modified distance calculation in Ref. [21] is introduced to enhance the calculation of IGD-NS indicator. Compared with IGD-NS, the enhanced IGD-NS (IGD-NS+) can not only comprehensively measure population but also have Pareto compliant.
- 2)
A reference point adaptation strategy is proposed for many-objective optimization. This method can adjust the distribution of reference points to adapt different types of Pareto fronts.
- 3)
The effectiveness of the proposed reference point adaptation method is assessed via comparisons with two reference point adaptation methods. It demonstrates that the proposed method has good performance for adapting different shapes of Pareto fronts.
The remainder of this paper is organized as follows. In section 2, related work of this paper is presented. The details of the proposed algorithm EIEA are described in section 3. The empirical results of EIEA compared with several representative MaOEAs are presented in section 4. Finally, section 5 concludes this paper.
Section snippets
Related work
Firstly, this section presents the related indicators and the modified distance calculation. Then, the existing reference point adaptation methods are illustrated.
The enhanced IGD-NS
IGD-NS can comprehensively measure a population whether has good convergence and diversity whereas it is Pareto non-compliant. To address this issue, equation (3) in section 2.1 is introduced into IGD-NS indicator to enhance the distance calculation of IGD-NS, namely IGD-NS+, which can be formulated as follows:where R is the set of reference point that are uniformly sampled on the PF, P is the nondominated solution set and is the
Experimental results and analysis
In this section, the experimental settings are first given. Then, the proposed EIEA is compared with several representative MaOEAs designed for solving MaOPs, namely, NSGA-III [13], RVEA [12], MOMBI-II [16] and AR-MOEA [17]. Also, the effectiveness of the proposed reference point adaptation method is assessed via comparisons with two reference point adaptation methods, namely, those in AR-MOEA [17] and the modified method of the proposed method in this paper. Finally, comparisons of running
Conclusion
This paper proposes an enhanced-indicator based many-objective evolutionary algorithm with adaptive reference point, termed EIEA, for solving MaOPs with different shapes of Pareto fronts. The EIEA proposes a reference point adaptation method to solve various shapes of Pareto fronts. The adaptation method introduces the evolutionary approach to adapt reference points, and the method also adopts hyperplane to adjust the distribution of reference points. In addition, the IGD-NS indicator is
CRediT authorship contribution statement
Junhua Li: Writing - original draft, Conceptualization. Guoyu Chen: Writing - review & editing. Ming Li: Validation, Project administration. Hao Chen: Validation, Project administration.
Acknowledgment
This work is supported by the National Natural Science Foundation of China (No.61440049, No.61866025 and No.61866026), the Natural Science Foundation of Jiangxi Province (NO.20181BAB202025), and the Superiority Science and Technology Innovation Team Program of Jiangxi Province (NO. 20181BCB24008).
References (35)
- et al.
Multiobjective evolutionary algorithms: a survey of the state of the art
Swarm and Evol. Comput.
(2011) - et al.
A novel multi-objective immune algorithm with a decomposition-based clonal selection
Appl. Soft Comput.
(2019) - et al.
Comparison between MOEA/D and NSGA-III on a set of novel many and multi-objective benchmark problems with challenging difficulties
Swarm and Evol. Comput.
(2019) - et al.
A survey of multiobjective evolutionary algorithms based on decomposition
IEEE Trans. Evol. Comput.
(2016) - et al.
Many-Objective evolutionary algorithms
ACM Comput. Surv.
(2015) - et al.
An investigation on preference order ranking scheme for multiobjective evolutionary optimization
IEEE Trans. Evol. Comput.
(2007) - et al.
Fuzzy-Based pareto optimality for many-objective evolutionary algorithms
IEEE Trans. Evol. Comput.
(2014) - et al.
A new dominance relation-based evolutionary algorithm for many-objective optimization
IEEE Trans. Evol. Comput.
(2016) - et al.
A knee point-driven evolutionary algorithm for many-objective optimization
IEEE Trans. Evol. Comput.
(2015) - et al.
An evolutionary many-objective optimization algorithm based on dominance and decomposition
IEEE Trans. Evol. Comput.
(2015)
MOEA/D: a multiobjective evolutionary algorithm based on decomposition
IEEE Trans. Evol. Comput.
A reference vector guided evolutionary algorithm for many-objective optimization
IEEE Trans. Evol. Comput.
An evolutionary many-objective optimization algorithm using reference-point-based nondominated sorting approach, part i: solving problems with box constraints
IEEE Trans. Evol. Comput.
Evolutionary many-objective optimization based on dynamical decomposition
IEEE Trans. Evol. Comput.
HypE: an algorithm for fast hypervolume-based many-objective optimization
Evol. Comput.
Improved metaheuristic based on the R2 indicator for many-objective optimization
An indicator-based multiobjective evolutionary algorithm with reference point adaptation for better versatility
IEEE Trans. Evol. Comput.
Cited by (17)
A dual distance dominance based evolutionary algorithm with selection-replacement operator for many-objective optimization
2024, Expert Systems with ApplicationsA survey of meta-heuristic algorithms in optimization of space scale expansion
2024, Swarm and Evolutionary ComputationAn adaptive two-stage evolutionary algorithm for large-scale continuous multi-objective optimization
2023, Swarm and Evolutionary ComputationProperty of decision variables-inspired location strategy for multiobjective optimization
2023, Swarm and Evolutionary ComputationCitation Excerpt :Compared with traditional algorithms, MOEAs are not limited by the mathematical properties of the objective functions and are not easy to fall into local optimum, which make them have obvious advantages in dealing with non-convex, non-differentiable, noisy and multimodal problems. According to different selection mechanisms, MOEAs can be divided into the following there categories: dominance relationship-based MOEAs [9–11], performance indicator-based MOEAs [12–14] and decomposition-based MOEAs [15–17]. Nevertheless, they work with a population, which results in slow convergence speed to the PF.
A reference vector based multiobjective evolutionary algorithm with Q-learning for operator adaptation
2023, Swarm and Evolutionary ComputationCitation Excerpt :The indicator-based MOEAs belong to the second category, using the performance indicator as the criteria for environmental selection, density estimation, archive update, and mating selection. The famous performance indicators include hypervolume indicator (HV), generational distance (GD), inverse generational distance (IGD), and shift based density estimator (SDE) [12–14]. HV considers the distance between the objective vector and the reference point, while GD and IGD investigate the distance between the objective vector and the point in the PF.
A survey of artificial immune algorithms for multi-objective optimization
2022, NeurocomputingCitation Excerpt :For example, a hybrid framework was designed in HEIA [39], which involves two evolutionary strategies that can complement with each other. The first one adopts the simulated binary crossover (SBX) [40–42] followed by polynomial-based mutation (PM) [43–45] that is validated to perform well on simple MOPs with independent decision variables. The second one applies DE crossover and PM, which is effective for solving some complicated MOPs with variable linkages.