Practical Model of Genetic Programming’s Performance on Rational Symbolic Regression Problems

Graff, Mario; Poli, Riccardo

doi:10.1007/978-3-540-78671-9_11

Cited by 7 publications

(15 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2 and 3 as a whole, we can see a fairly clear overall picture of the behaviour of GP systems over the class of Boolean problems. 6 In particular we find that behaviours tend to be relatively little influenced by the choice of crossover and mutation rates, while changes in the selection and reproduction schemes can give substantial performance differences. This in turn suggests that adjusting crossover and mutation rates is particularly useful if one has already found a good GP system for a problem (or group of problems) and now wants to get the maximum out of that system.…”

Section: Taxonomiesmentioning

confidence: 78%

“…So, we decided against this approach and instead adopted a more complex, but also more general technique which consists in creating the c vectors using the coefficients of a model of GP behaviour. In particular, we will use a version of the performance model originally proposed in [6]. We review it in the next section.…”

Section: Creating Gp's Taxonomiesmentioning

confidence: 99%

“…The model proposed in [6] is a practical model of GP which has been specifically designed to focus on performance. Its applicability was tested on the class of rational symbolic regression problems with excellent results using as a performance measure the best fitness recorded in a run.…”

Section: Performance Modelmentioning

confidence: 99%

“…This is a point where our approach deviates from [6] where a GA and ordinary least squares were used for this purposes, while here we use a much simpler, yet effective approach. In the next section we provide details on the model identification techniques we used.…”

Section: Performance Modelmentioning

confidence: 99%

“…However, our performance signatures represent the behaviour of each system across a class of problems, not the behaviour of multiple versions of an algorithm over one problem. Also, our signatures are not the measurement of performance over multiple cases: they are obtained thorough a process which involves the instantiation of a mathematical models of GP's performance recently introduced in [6]. It is then this model which provides the signature for each algorithm.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Automatic Creation of Taxonomies of Genetic Programming Systems

Graff

Poli

2009

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Abstract. A few attempts to create taxonomies in evolutionary computation have been made. These either group algorithms or group problems on the basis of their similarities. Similarity is typically evaluated by manually analysing algorithms/problems to identify key characteristics that are then used as a basis to form the groups of a taxonomy. This task is not only very tedious but it is also rather subjective. As a consequence the resulting taxonomies lack universality and are sometimes even questionable. In this paper we present a new and powerful approach to the construction of taxonomies and we apply it to Genetic Programming (GP). Only one manually constructed taxonomy of problems has been proposed in GP before, while no GP algorithm taxonomy has ever been suggested. Our approach is entirely automated and objective. We apply it to the problem of grouping GP systems with their associated parameter settings. We do this on the basis of performance signatures which represent the behaviour of each system across a class of problems. These signatures are obtained thorough a process which involves the instantiation of models of GP's performance. We test the method on a large class of Boolean induction problems.

show abstract

Section: Taxonomiesmentioning

confidence: 78%

Section: Creating Gp's Taxonomiesmentioning

confidence: 99%

Section: Performance Modelmentioning

confidence: 99%

Section: Performance Modelmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Automatic Creation of Taxonomies of Genetic Programming Systems

Graff

Poli

2009

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

show abstract

Performance Models for Evolutionary Program Induction Algorithms Based on Problem Difficulty Indicators

Graff

Poli

2011

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Theoretical results in genetic programming: the next ten years?

Poli

Vanneschi

Langdon

et al. 2010

Genet Program Evolvable Mach

Self Cite

View full text Add to dashboard Cite

We consider the theoretical results in GP so far and prospective areas for the future. We begin by reviewing the state of the art in genetic programming (GP) theory including: schema theories, Markov chain models, the distribution of functionality in program search spaces, the problem of bloat, the applicability of the no-free-lunch theory to GP, and how we can estimate the difficulty of problems before actually running the system. We then look at how each of these areas might develop in the next decade, considering also new possible avenues for theory, the challenges ahead and the open issues.

show abstract

Practical Model of Genetic Programming’s Performance on Rational Symbolic Regression Problems

Cited by 7 publications

References 4 publications

Automatic Creation of Taxonomies of Genetic Programming Systems

Automatic Creation of Taxonomies of Genetic Programming Systems

Performance Models for Evolutionary Program Induction Algorithms Based on Problem Difficulty Indicators

Theoretical results in genetic programming: the next ten years?

Contact Info

Product

Resources

About