AREA: An adaptive reference-set based evolutionary algorithm for multiobjective optimisation

Abstract

Population-based evolutionary algorithms have great potential to handle multiobjective optimisation problems. However, these algorithms depends largely on problem characteristics, and there is a need to improve their performance for a wider range of problems. References, which are often specified by the decision maker's preference in different forms, are a very effective method to improve the performance of algorithms but have not been fully explored in literature. This paper proposes a novel framework for effective use of references to strengthen algorithms. This framework considers references as search targets which can be adjusted based on the information collected during the search. The proposed framework is combined with new strategies, such as reference adaptation and adaptive local mating, to solve different types of problems. The proposed algorithm is compared with state of the arts on a wide range of problems with diverse characteristics. The comparison and extensive sensitivity analysis demonstrate that the proposed algorithm is competitive and robust across different types of problems studied in this paper.

Benchmarking Real-World Many-Objective Problems: A Problem Suite With Baseline Results

“…As a result of these intricate challenges, managing MaOPs is a difficult endeavor. Therefore, various advanced MOEA approaches, including Pareto-dominance-based [18]- [21], indicator-based [22]- [24], decomposition-based [25]- [28], and hybrid MOEAs [29]- [31], have been presented in the literature to address the MaOPs. However, the performance evaluation of most MOEAs is conducted using artificial test problems.…”

“…In this study, we have selected eight recent state-of-theart algorithms designed to solve MaOPs. The algorithms selected in this study belong to different selection criteria proposed for MOEAs.The algorithms that have been chosen for this study are MultiGPO [19], Pi-MOEA [18], MaOEA-IBP [22], MOEAD-URAW [26], AREA [25], PeEA [47], TSNSGAII [30], and VMEF [29]. MultiGPO The primary aim of this framework is to enhance both convergence (the ability to approach the optimal solution) and diversity (the representation of a wide range of solutions) in the optimization process.…”

“…where L, B, D, T, V k , C B correspond to six decision variables (x 1 , x 2 , x 3 , x 4 , x 5 , x 6 ), respectively. The range of the six variables is as follows: L ∈ [150, 274.32], B ∈ [20, 32.31], D ∈ [13,25], T ∈ [10, 11.71], V k ∈ [14,18], and C B ∈ [0.63, 0.75]…”

“…Additionally, MOEAD-URAW algorithms adapt weightings to accommodate different shapes of Pareto Fronts. 5) AREA[25]: The AREA method differs from other decomposition-based algorithms in using reference points as population targets instead of references for spatial decomposition. This approach allows the solution set's estimated Pareto Front (PF) to adapt and modify the reference set, making the population robust and capable of handling various PF geometries.…”

Many-Objective Real-World Engineering Problems: A Comparative Study of State-of-the-Art Algorithms

“…have been developed to address MaOPs. However, when dealing with MaOPs, algorithms encounter three main challenges [18]. First, as the number of objectives increases, the nondominated solutions increase, resulting in the weakened selection pressure toward PF.…”