Integration of discrete and continuous multi-objective optimization benchmark problems
| Reference No. | 2024c005 |
|---|---|
| Type/Category | Openings at any time-Short-term Joint Research |
| Title of Research Project | Integration of discrete and continuous multi-objective optimization benchmark problems |
| Principal Investigator | Shoichiro Tanaka（The university of Fukuchiyama・Assistant professor） |
| Research Period |
January 20,2025. -
January 22,2025. |
| Keyword(s) of Research Fields | Multi-objective optimization, Benchmark, Topological data analysis, Landscape analysis |
| Abstract for Research Report |
In industry, it is common to encounter multi-objective optimization problems with conflicting objectives, such as maximizing performance while minimizing costs. Evolutionary multi-objective optimization (EMO) algorithms have gained attention as effective methods for tackling these problems. But there is no all-round algorithm that solves all problems efficiently. The interaction between the problem and EMO algorithms has a significant impact on search performance. Therefore, practitioners must carefully select an algorithm according to the targeting problem. To benchmark EMO algorithms effectively, it is essential to design artificial problems with tunable properties. In the discrete multi-objective optimization domain, benchmark problems like the ρMNK-landscape enable tuning of both the solution space and objective space, providing valuable insights into how problem characteristics affect algorithm performance. In contrast, benchmark problems in the continuous domain primarily focus on the objective space, with the influence of solution space being less thoroughly investigated. This underscores the need to transfer knowledge from the discrete to the continuous domain. Moreover, it is crucial to determine what type of test suite (benchmark problem set) can be designed to evaluate the performance of EMO algorithms in realistic industrial scenarios. This collaborative research bridges the gap between discrete and continuous optimization by bringing experts in both fields and practitioners from industry. |
| Organizing Committee Members (Workshop) Participants (Short-term Joint Usage) |
Shoichiro Tanaka（The University of Fukuchiyama・Assistant professor） Daisuke Sakurai（Fujitsu Limited・Research director） Naoki Hamada（KLab Inc.・Researcher） Naoyuki Kamiyama（Kyushu University・Professor） |