A comparative analysis of two matheuristics by means of merged local optima networks

Blum, Christian; Ochoa, Gabriela

doi:10.1016/j.ejor.2020.08.008

“…An alternative would be to directly compare trajectories. This is much harder to do, one potential approach would be to map trajectories to local optima networks (as done in [4]), and then compare the trajectories within these networks.…”

Section: Limitations and Future Workmentioning

confidence: 99%

Evolving Continuous Optimisers from Scratch

Lones

2021

Preprint

View full text Add to dashboard Cite

This work uses genetic programming to explore the space of continuous optimisers, with the goal of discovering novel ways of doing optimisation. In order to keep the search space broad, the optimisers are evolved from scratch using Push, a Turing-complete, general-purpose, language. The resulting optimisers are found to be diverse, and explore their optimisation landscapes using a variety of interesting, and sometimes unusual, strategies. Significantly, when applied to problems that were not seen during training, many of the evolved optimisers generalise well, and often outperform existing optimisers. This supports the idea that novel and effective forms of optimisation can be discovered in an automated manner. This paper also shows that pools of evolved optimisers can be hybridised to further increase their generality, leading to optimisers that perform robustly over a broad variety of problem types and sizes.

show abstract

“…An alternative would be to directly compare trajectories. This is much harder to do, but one potential approach would be to map trajectories to local optima networks (as done in [4]), and then compare the trajectories within these networks.…”

Section: Limitations and Future Workmentioning

confidence: 99%

Evolving continuous optimisers from scratch

Lones

¹

2021

Genet Program Evolvable Mach

6

0

View full text Add to dashboard Cite

This work uses genetic programming to explore the space of continuous optimisers, with the goal of discovering novel ways of doing optimisation. In order to keep the search space broad, the optimisers are evolved from scratch using Push, a Turing-complete, general-purpose, language. The resulting optimisers are found to be diverse, and explore their optimisation landscapes using a variety of interesting, and sometimes unusual, strategies. Significantly, when applied to problems that were not seen during training, many of the evolved optimisers generalise well, and often outperform existing optimisers. This supports the idea that novel and effective forms of optimisation can be discovered in an automated manner. This paper also shows that pools of evolved optimisers can be hybridised to further increase their generality, leading to optimisers that perform robustly over a broad variety of problem types and sizes.

show abstract

Introduction to CMSA

Blum

2024

Computational Intelligence Methods and Applications

0

View full text Add to dashboard Cite

A comparative analysis of two matheuristics by means of merged local optima networks

Cited by 11 publications

References 43 publications

Evolving Continuous Optimisers from Scratch

Evolving Continuous Optimisers from Scratch

Evolving continuous optimisers from scratch

Introduction to CMSA

Contact Info

Product

Resources

About