Study reveals inefficiencies in metaphor-based metaheuristics, calls for automation. Researchers recommend automatic design for metaheuristics to improve outcomes.

In an effort to discourage the inefficient and outdated practice of manually creating and configuring new metaheuristic optimization algorithms, a team of researchers at IRIDIA, a leading artificial intelligence laboratory at the Université Libre de Bruxelles, has conducted a thorough review of existing literature. Their findings, published in Intelligent Computing, highlight the advantages of automatic approaches to designing metaheuristics, particularly when compared to the numerous redundant and often far-fetched metaphor-based methods.

The team's review not only looks back at the history of metaheuristics, but also looks ahead to the future. They recommend increasing the use of modular metaheuristic software frameworks and automatic configuration tools to eliminate the wasteful trial-and-error process associated with manual design methods, ultimately leading to more successful outcomes. This is crucial in the pursuit of optimizing the search for metaheuristic optimization algorithms, which are used to solve computational problems in a variety of contexts, such as machine learning.

Some of the most well-known and effective metaheuristics include evolutionary computation, simulated annealing, and ant colony optimization, all of which draw inspiration from natural processes. However, as the review highlights, the field of metaheuristics is cluttered with similarly nature-inspired and metaphorically named contributions that may not offer any new or useful techniques.

The authors of the review express concern over the large number of supposedly groundbreaking metaphor-based metaheuristics that only serve to add confusion and create unnecessary terminology, rather than offering any truly novel methods. They cite examples of algorithms based on the behavior of wolves, cuckoos, zombies, reincarnation, and even "intelligent" water drops that know how to move towards bodies of water.

While the "metaphor rush" remains an ongoing issue, the real problem lies in the reliance on inspiration rather than scientific design principles. The authors urge researchers to instead focus on automatic design, which requires a metaheuristic software framework to provide a design space and an automatic configuration tool to test different combinations of components. They note that modern frameworks, such as ParadisEO, HeuristicLab, jMetal, and EMILI, are continuously evolving and offer great flexibility and modularity.

Looking to the future, the review suggests exploring other paths for productive research, such as modeling metaheuristics in greater detail and developing statistical tools for benchmarking, which would provide a more solid foundation for the field. By embracing automatic design and utilizing advanced software frameworks, the field of metaheuristics can continue to evolve and achieve even greater success in solving complex computational problems.

Ann Castro
Ann Castro Author
Ann Castro carries a total of 7 years experience in the healthcare domain. She owns a Master’s of Medicine Degree. She bagged numerous awards by contributing in the medical field with her ground-breaking notions. Ann has developed her own style of working and known for accuracy in her work. She loves trekking. She visits new places whenever she gets free time.