Proceedings of the Companion Publication of the 2014 Annual Conference on Genetic and Evolutionary Computation 2014
DOI: 10.1145/2598394.2605684
|View full text |Cite
|
Sign up to set email alerts
|

A quantitative analysis of the simplification genetic operator

Abstract: The simplification function was introduced to PushGP as a tool to reduce the sizes of evolved programs in final reports. While previous work suggests that simplification could reduce the sizes significantly, nothing has been done to study its impacts on the evolution of Push programs. In this paper, we show the impact of simplification as a genetic operator. By conducting test runs on the U.S. change problem, we show that using simplification operator with PushGP, lexicase selection and ULTRA could increase th… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1

Citation Types

0
2
0

Year Published

2015
2015
2020
2020

Publication Types

Select...
4
1

Relationship

0
5

Authors

Journals

citations
Cited by 9 publications
(2 citation statements)
references
References 5 publications
0
2
0
Order By: Relevance
“…While it can be applied to any Push program, it has typically been used only aer the completion of GP runs, to make solution programs easier to understand without changing their behavior [29]. It has also been tried as a growth-control genetic operator during GP runs, with mixed success [35]. Here we consider only post-run simplication, and we focus only on its eects on code size and program generalization.…”
Section: Introductionmentioning
confidence: 99%
“…While it can be applied to any Push program, it has typically been used only aer the completion of GP runs, to make solution programs easier to understand without changing their behavior [29]. It has also been tried as a growth-control genetic operator during GP runs, with mixed success [35]. Here we consider only post-run simplication, and we focus only on its eects on code size and program generalization.…”
Section: Introductionmentioning
confidence: 99%
“…Three of these (K Landscapes, ORDERTREE, and Pagie-1 regression) are taken from recent benchmark suggestions [24]. The U.S. Change problem [26] has proven to be an interesting problem in evolutionary program synthesis [7], and sine regression is the example problem used in earlier work on the impact of root and non-root parents [10]. All experiments were run using a copy of ECJ 21 1 that we modified to support crossover bias.…”
Section: Methodsmentioning
confidence: 99%