.png){kind=icon}
.png){kind=icon}
Return
Linear prediction evolution algorithm: a simplest evolutionary optimizer
.png){kind=icon}
PRE
AI
DOI：10.1007/s12293-021-00340-x.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The prediction-based evolutionary algorithms are a recently developed branch of metaheuristic algorithms. The most notable feature of this kind of algorithms is the use of a certain prediction model to develop their reproduction operators for evolution. The linear least square fitting model, as a simplest and most widely used statistic model, is first introduced to construct a linear prediction evolution algorithm (LPE) in this paper. Firstly, the proposed LPE randomly selects three individuals from three consecutive populations, respectively, and then fits a line on each dimension of the three individuals by using the linear least square fitting model. Finally, LPE regards the line expression as its reproduction operator to generate the offspring individuals. LPE algorithm does not have any control parameters except for a population size. Its reproduction operator based on the linear least square fitting model holds solid mathematical foundation without any empirical coefficients, and is theoretically proven to be adaptive to the variation of population regions. The effectiveness of the proposed LPE is validated on CEC2014, CEC2017 benchmark functions and a comprehensive set of seven engineering design problems. The comparison experiments indicate that LPE is a competitive optimizer compared with other state-of-the-art algorithms.
Keywords:
Engineering design problems
Linear least square fitting model
Prediction-based evolutionary algorithms
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
2.3
Papers:
453
Citations:
718