Exploring position independent initialisation in grammatical evolution

Fagan, David; Fenton, Michael; O’Neill, Michael

doi:10.1109/cec.2016.7748331

“…Whereas RHH generates pairs of trees at a range of depths, PI Grow eschews the combination of full and grow derivations and generates individuals at a range of depths where at least one branch of the derivation tree is forced to the given depth. Furthermore, to combat the leftmost derivation tendencies of pre-fix or in-fix grammarbased mapping systems, PI Grow derives trees in a position independent manner by randomising the order of derivation of non-terminals [8]. This has the effect of reducing inherent biases which are intrinsic to grammar-based systems.…”

Section: Initialisation and Population Seedingmentioning

confidence: 99%

“…At each depth, derivation trees are created using "full" (every branch in the tree is forced to the given depth) and "grow" (no branch in the tree is forced to the depth and the tree is allowed to grow at random up to that given depth) derivation methods. While this technique is widely used, there have been concerns over its appropriateness in certain applications [25,14,8,27]. A single seed individual program is then added to the initialised population, resulting in the population instantly gaining a highly fit local optimum.…”

Section: Initialisation and Population Seedingmentioning

confidence: 99%

“…PI Grow + Seed Position Independent Grow (PI Grow) has been proposed as a viable alternative to RHH initialisation [8]. Whereas RHH generates pairs of trees at a range of depths, PI Grow eschews the combination of full and grow derivations and generates individuals at a range of depths where at least one branch of the derivation tree is forced to the given depth.…”

Section: Initialisation and Population Seedingmentioning

confidence: 99%

See 1 more Smart Citation

A search for improved performance in regular expressions

Cody-Kenny¹,

Fenton²,

Ronayne³

et al. 2017

Proceedings of the Genetic and Evolutionary Computation Conference

Self Cite

View full text Add to dashboard Cite

The primary aim of automated performance improvement is to reduce the running time of programs while maintaining (or improving on) functionality. In this paper, Genetic Programming is used to find performance improvements in regular expressions for an array of target programs, representing the first application of automated software improvement for run-time performance in the Regular Expression language. This particular problem is interesting as there may be many possible alternative regular expressions which perform the same task while exhibiting subtle differences in performance. A benchmark suite of candidate regular expressions is proposed for improvement. We show that the application of Genetic Programming techniques can result in performance improvements in all cases.As we start evolution from a known good regular expression, diversity is critical in escaping the local optima of the seed expression. In order to understand diversity during evolution we compare an initial population consisting of only seed programs with a population initialised using a combination of a single seed individual with individuals generated using PI Grow and Ramped-half-and-half initialisation mechanisms. * Brendan Cody-Kenny, Michael Fenton, and Michael O'Neill are with the Natural Computing Research and Applications Group (NCRA) in University College Dublin, Ireland. Email: Brendan.Cody-Kenny@ucd.ie, MichaelFenton1@gmail.com, M.ONeill@ucd.ie. † Adrian Ronayne, Eoghan Considine and Thomas McGuire, are with the Fidelity Center for Applied Technology (FCAT) in Fidelity Investments, Ireland.Program execution time can be difficult to measure and reason about for a number of reasons [33]. Reading through program source code can only partially indicate expected

show abstract

“…For this problem we used a modified version of the grammar given in [13], and [4], with the only modification of increasing constant and variable token ratio as the expression nesting gets deeper. We used the root mean squared error as fitness function which is the accepted practice for this problem.…”

Section: Keijzer-6mentioning

confidence: 99%

Parallel and in-process compilation of individuals for genetic programming on GPU

Ayral

¹

,

Albayrak

²

2017

Preprint

0

View full text Add to dashboard Cite

Three approaches to implement genetic programming on GPU hardware are compilation, interpretation and direct generation of machine code. The compiled approach is known to have a prohibitive overhead compared to other two. This paper investigates methods to accelerate compilation of individuals for genetic programming on GPU hardware. We apply in-process compilation to minimize the compilation overhead at each generation; and we investigate ways to parallelize in-process compilation. In-process compilation doesn't lend itself to trivial parallelization with threads; we propose a multiprocess parallelization using memory sharing and operating systems interprocess communication primitives. With parallelized compilation we achieve further reductions on compilation overhead. Another contribution of this work is the code framework we built in C# for the experiments. The framework makes it possible to build arbitrary grammatical genetic programming experiments that run on GPU with minimal extra coding effort, and is available as open source. Three approaches to implement genetic programming on GPU hardware are compilation, interpretation and direct generation of machine code. The compiled approach is known to have a prohibitive overhead compared to other two. This paper investigates methods to accelerate compilation of individuals for genetic programming on GPU hardware. We apply in-process compilation to minimize the compilation overhead at each generation; and we investigate ways to parallelize in-process compilation. In-process compilation doesn't lend itself to trivial parallelization with threads; we propose a multiprocess parallelization using memory sharing and operating systems interprocess communication primitives. With parallelized compilation we achieve further reductions on compilation overhead. Another contribution of this work is the code framework we built in C# for the experiments. The framework makes it possible to build arbitrary grammatical genetic programming experiments that run on GPU with minimal extra coding effort, and is available as open source.

show abstract

“…For this problem we used a modified version of the grammar given in Nicolau and Fenton (2016), and Fagan et al (2016), with the only modification of increasing constant and variable token ratio as the expression nesting gets deeper. We used the root mean squared error as fitness function which is the accepted practice for this problem.…”

Section: Keijzer-6mentioning

confidence: 99%

Parallel and in-process compilation of individuals for genetic programming on GPU

Ayral

¹

,

Albayrak

²

2017

Preprint

0

View full text Add to dashboard Cite

Three approaches to implement genetic programming on GPU hardware are compilation, interpretation and direct generation of machine code. The compiled approach is known to have a prohibitive overhead compared to other two. This paper investigates methods to accelerate compilation of individuals for genetic programming on GPU hardware. We apply in-process compilation to minimize the compilation overhead at each generation; and we investigate ways to parallelize in-process compilation. In-process compilation doesn't lend itself to trivial parallelization with threads; we propose a multiprocess parallelization using memory sharing and operating systems interprocess communication primitives. With parallelized compilation we achieve further reductions on compilation overhead. Another contribution of this work is the code framework we built in C# for the experiments. The framework makes it possible to build arbitrary grammatical genetic programming experiments that run on GPU with minimal extra coding effort, and is available as open source. Three approaches to implement genetic programming on GPU hardware are compilation, interpretation and direct generation of machine code. The compiled approach is known to have a prohibitive overhead compared to other two. This paper investigates methods to accelerate compilation of individuals for genetic programming on GPU hardware. We apply in-process compilation to minimize the compilation overhead at each generation; and we investigate ways to parallelize in-process compilation. In-process compilation doesn't lend itself to trivial parallelization with threads; we propose a multiprocess parallelization using memory sharing and operating systems interprocess communication primitives. With parallelized compilation we achieve further reductions on compilation overhead. Another contribution of this work is the code framework we built in C# for the experiments. The framework makes it possible to build arbitrary grammatical genetic programming experiments that run on GPU with minimal extra coding effort, and is available as open source.

show abstract

Exploring position independent initialisation in grammatical evolution

Cited by 30 publications

References 7 publications

A search for improved performance in regular expressions

A search for improved performance in regular expressions

Parallel and in-process compilation of individuals for genetic programming on GPU

Parallel and in-process compilation of individuals for genetic programming on GPU

Contact Info

Product

Resources

About