Investigating the Properties of Indicators and an Evolutionary Many-Objective Algorithm Using Promising Regions

Jiawei Yuan,Hai-Lin Liu,Fangqing Gu,Qingfu Zhang,Zhaoshui He

DOI: https://doi.org/10.1109/tevc.2020.2999100

IF: 16.497

2021-02-01

IEEE Transactions on Evolutionary Computation

Abstract:This article investigates the properties of ratio and difference-based indicators under the Minkovsky distance and demonstrates that a ratio-based indicator with infinite norm is the best for solution evaluation among these indicators. Accordingly, a promising-region-based evolutionary many-objective algorithm with the ratio-based indicator is proposed. In our proposed algorithm, a promising region is identified in the objective space using the ratio-based indicator with infinite norm. Since the individuals outside the promising region are of poor quality, we can discard these solutions from the current population. To ensure the diversity of population, a strategy based on the parallel distance is introduced to select individuals in the promising region. In this strategy, all individuals in the promising region are projected vertically onto the normal plane so that crowded distances between them can be calculated. Afterward, two solutions with a smaller distance are selected from the candidate solutions each time, and the solution with the smaller indicator fitness value is removed from the current population. Empirical studies on various benchmark problems with 3–20 objectives show that the proposed algorithm performs competitively on all test problems. Compared with a number of other state-of-the-art evolutionary algorithms, the proposed algorithm is more robust on these problems with various Pareto fronts.

computer science, artificial intelligence, theory & methods

What problem does this paper attempt to address?

### Problems Addressed by the Paper

This paper primarily explores several key issues in multi-objective optimization (MOP) and proposes a new evolutionary algorithm to address these issues. Specifically:

1. **Indicator Analysis**:
- Investigated the properties of ratio-based and difference-based indicators under Minkowski distance.
- Found that the ratio-based indicator under the infinity norm outperforms other indicators in evaluating solution quality.

2. **Algorithm Design**:
- Proposed an evolutionary multi-objective algorithm based on promising regions (PREA).
- Identified promising regions in the current population using the ratio-based indicator under the infinity norm and selected high-quality solutions from these regions.
- Employed a parallel distance strategy to maintain the diversity of the final solution set.

3. **Experimental Validation**:
- Conducted extensive benchmark tests on Pareto fronts of different shapes to demonstrate the effectiveness and robustness of the proposed algorithm.

### Main Contributions

1. **Theoretical Analysis**:
- Analyzed the differences in handling outliers and protecting boundary points in the population using different indicators.
- Proved that the ratio-based indicator under the infinity norm can effectively eliminate outliers and protect boundary points on the Pareto front.

2. **Algorithm Innovation**:
- Utilized the ratio-based indicator under the infinity norm to identify promising regions and remove inferior individuals from the current population.
- Introduced a diversity maintenance mechanism based on parallel distance to ensure the diversity of the final solution set.

3. **Experimental Results**:
- Experimental results show that the proposed algorithm exhibits better robustness on Pareto front problems of different shapes.
- Compared to several existing advanced algorithms, this algorithm performs superiorly.

Through the above research, the paper aims to improve the performance and robustness of evolutionary multi-objective optimization algorithms in handling high-dimensional multi-objective optimization problems.