Evolving search spaces to emphasize the performance difference of real-coded crossovers using genetic programming

Shirakawa, Shinichi; Yata, Noriko; Nagao, Tomoharu

doi:10.1109/cec.2010.5586065

Cited by 4 publications

(2 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our current work is focused on developing the key components of the methodology (evolved instances, feature sets) for a number of broad classes of optimization problems [13], which will then become available as tools for researchers. The generalization of the approach to consider not just algorithm selection, but parameter selection (for example, evolutionary operator selection [26] is a natural extension for the evolutionary computation community.…”

Section: Discussionmentioning

confidence: 99%

Measuring algorithm footprints in instance space

Smith‐Miles

Tan

2012

2012 IEEE Congress on Evolutionary Computation

View full text Add to dashboard Cite

This paper proposes a new methodology to determine the relative performance of optimization algorithms across various classes of instances. Rather than reporting performance based on a chosen test set of benchmark instances, we aim to develop metrics for an algorithm's performance generalized across a diverse set of instances. Instances are summarized by a set of features that correlate with difficulty, and we propose methods for visualizing instances and algorithm performance in this high-dimensional feature space. The footprint of an algorithm is where good performance can be expected, and we propose new metrics to measure the relative size of an algorithm's footprint in instance space. The methodology is demonstrated using the Traveling Salesman Problem as a case study.

show abstract

Section: Discussionmentioning

confidence: 99%

Measuring algorithm footprints in instance space

Smith‐Miles

Tan

2012

2012 IEEE Congress on Evolutionary Computation

View full text Add to dashboard Cite

show abstract

“…Genetic programming was used to evolve problem landscapes [30] to analyze the performance of existing EAs. Thereafter, the test suite [30] was extended to MOOPs using the combination of three types of real-coded crossovers [53].…”

Section: Related Workmentioning

confidence: 99%

An Open Framework for Constructing Continuous Optimization Problems

Nguyễn

Zeng

et al. 2019

IEEE Trans. Cybern.

View full text Add to dashboard Cite

Many artificial benchmark problems have been proposed for different kinds of continuous optimization, e.g., global optimization, multimodal optimization, multiobjective optimization, dynamic optimization, and constrained optimization. However, there is no unified framework for constructing these types of problems and possible properties of many problems are not fully tunable. This will cause difficulties for researchers to analyze strengths and weaknesses of an algorithm. To address these issues, this paper proposes a simple and intuitive framework, which is able to construct different kinds of problems for continuous optimization. The framework utilizes the k-d tree to partition the search space and sets a certain number of simple functions in each subspace. The framework is implemented into global/multimodal optimization, dynamic single objective optimization, multiobjective optimization, and dynamic multiobjective optimization, respectively. Properties of the proposed framework are discussed and verified with traditional evolutionary algorithms.

show abstract