Genetic programming needs better benchmarks

McDermott, James; White, David R.; Luke, Sean; Manzoni, Luca; Castelli, Mauro; Vanneschi, Leonardo; Jaśkowski, Wojciech; Krawiec, Krzysztof; Harper, Robin; Jong, Kenneth De; O’Reilly, Una-May

doi:10.1145/2330163.2330273

Cited by 235 publications

(183 citation statements)

References 60 publications

Supporting

Mentioning

182

Contrasting

Order By: Relevance

“…The proposed NS-based GP was compared with recently published results on three benchmark problems that are currently suggested for GP evaluation within the community [32]. NS shows a consistent trend, it achieved quite bad performance on easy problems, and performs substantially better on difficult ones, results that are similar to those published in [20], [21].…”

Section: Discussionmentioning

confidence: 54%

“…The proposed NS-GP-R algorithm is evaluated on three benchmark problems, suggested by [32] and proposed in [24]. Moreover, for comparative purposes, the algorithm is compared to the results published in [24], that use a standard GP, hereafter referred to as SGP, and a GP with the Semantic Similarity-based Crossover (SSC) also proposed in [24].…”

Section: Methodsmentioning

confidence: 99%

“…For now, it suffices to say that behaviors are a more general concept than semantics, that can provide a simplified view of what a program is actually doing. Therefore, this work proposes a behavioral descriptor for symbolic regression problems, and results are compared with published results on benchmark tests [32], [24]. The results are consistent with those published in [20], [21], showing that NS performs better when problem difficulty increases, which is what would be expected from a search process that starts with low-quality behaviors and must necessarily explore novel regions in the search space, some of which will contain high-quality solutions; i.e.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Searching for novel regression functions

Martínez

Naredo

Trujillo

et al. 2013

2013 IEEE Congress on Evolutionary Computation

View full text Add to dashboard Cite

Abstract-The objective function is the core element in most search algorithms that are used to solve engineering and scientific problems, referred to as the fitness function in evolutionary computation. Some researchers have attempted to bridge this difference by reducing the need for an explicit fitness function. A noteworthy example is the novelty search (NS) algorithm, that substitutes fitness with a measure of uniqueness, or novelty, that each individual introduces into the search. NS employs the concept of behavioral space, where each individual is described by a domain-specific descriptor that captures the main features of an individual's performance. However, defining a behavioral descriptor is not trivial, and most works with NS have focused on robotics. This paper is an extension of recent attempts to expand the application domain of NS. In particular, it represents the first attempt to apply NS on symbolic regression with genetic programming (GP). The relationship between the proposed NS algorithm and recent semantics-based GP algorithms is explored. Results are encouraging and consistent with recent findings, where NS achieves below average performance on easy problems, and achieves very good performance on hard problems. In summary, this paper presents the first attempt to apply NS on symbolic regression, a continuation of recent research devoted at extending the domain of competence for behavior-based search.

show abstract

Section: Discussionmentioning

confidence: 54%

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Searching for novel regression functions

Martínez

Naredo

Trujillo

et al. 2013

2013 IEEE Congress on Evolutionary Computation

View full text Add to dashboard Cite

show abstract

“…All of the experiments are investigated using the six bit parity and the Pagie1 [16] symbolic regression tasks. The parity task uses AND NAND OR and NOR node functions 2 and the fitness is calculated as the number of incorrect outputs produced when all possible inputs are swept.…”

Section: Methodsmentioning

confidence: 99%

Cartesian Genetic Programming: Why No Bloat?

Turner

Miller

2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. For many years now it has been known that Cartesian Genetic Programming (CGP) does not exhibit program bloat. Two possible explanations have been proposed in the literature: neutral genetic drift and length bias. This paper empirically disproves both of these and thus, reopens the question as to why CGP does not suffer from bloat. It has also been shown for CGP that using a very large number of nodes considerably increases the effectiveness of the search. This paper also proposes a new explanation as to why this may be the case.

show abstract

“…This problem has also been used for benchmarking (Harper, 2012), and has been recommended as a replacement for "toy" problems such as symbolic regression of the quartic polynomial (McDermott et al, 2012;White et al, 2013).…”

Section: Methodsmentioning

confidence: 99%

Uniform Linear Transformation with Repair and Alternation in Genetic Programming

Spector

Helmuth

2014

Genetic and Evolutionary Computation

View full text Add to dashboard Cite

Several genetic programming researchers have argued for the utility of genetic operators that act uniformly. By "act uniformly" we mean two specific things: that the probability of an inherited program component being modified during inheritance is independent of the size and shape of the parent programs beyond the component in question; and that pairs of parents are combined in ways that allow arbitrary combinations of components from each parent to appear in the child. Uniform operators described in previous work have had limited utility, however, because of a mismatch between the relevant notions of uniformity and the hierarchical structure and variable sizes of many genetic programming representations. In this chapter we describe a new genetic operator, ULTRA, which incorporates aspects of both mutation and crossover and acts approximately uniformly across programs of variable sizes and structures. ULTRA treats hierarchical programs as linear sequences and includes a repair step to ensure that syntax constraints are satisfied after variation. We show that on the drug bioavailability and Pagie-1 benchmark problems ULTRA produces significant improvements both in problem-solving power and in program size relative to standard operators. Experiments with factorial regression and with the boolean 6-multiplexer problem demonstrate that ULTRA can manipulate programs that make use of hierarchical structure, but also that it is not always beneficial. The demonstrations evolve programs in the Push programming language, which makes repair particularly simple, but versions of the technique should be applicable in other genetic programming systems as well.

show abstract

Genetic programming needs better benchmarks

Cited by 235 publications

References 60 publications

Searching for novel regression functions

Searching for novel regression functions

Cartesian Genetic Programming: Why No Bloat?

Uniform Linear Transformation with Repair and Alternation in Genetic Programming

Contact Info

Product

Resources

About