Genetic Improvement using Higher Order Mutation

Yue Jia,Fan Wu,Mark Harman,Jens Krinke

DOI: https://doi.org/10.1145/2739482.2768417

2015-01-01

Abstract:This paper presents a brief outline of a higher-order mutation-based framework for Genetic Improvement (GI). We argue that search-based higher-order mutation testing can be used to implement a form of genetic programming (GP) to increase the search granularity and testability of GI.

What problem does this paper attempt to address?

Parameter estimation for time-varying system based on improved genetic algorithm

Yuncan Xue,Qiwen Yang,Jixin Qian

DOI: https://doi.org/10.1109/IECON.2002.1185280

2002-01-01

Abstract:In this paper, the authors first analyze the drawback of genetic algorithm with traditional mutation operator in preventing GA from converging prematurely, then study its influence on the diversity of genes in the same locus, and point out that the traditional mutation, to some extent, can result in premature convergence of genes in the same locus. This will cause the loss of critical alleles. Inspired by digital technique, we import two kinds of Boolean operation into GA to develop a novel mutation operator and discuss its contribution to prevent the loss of critical alleles. The experimental results of function optimization show that the improved mutation operator can prevent premature convergence effectively, and can provide a wide range of selecting control parameters for GA. Then we present a new parameter estimation algorithm based on the improved GA. Simulation experiments have been carried out and their results show that this estimation algorithm has good tracking ability and is not sensitive to the noise.
Improved Genetic Operator for Genetic Algorithm

Lin Feng,Yang Qi-wen

DOI: https://doi.org/10.1631/bf02839485

2002-01-01

Journal of Zhejiang University SCIENCE A

Abstract:The mutation operator has been seldom improved because researchers hardly suspect its ability to prevent genetic algorithm (GA) from converging prematurely. Due to its importance to GA, the authors of this paper study its influence on the diversity of genes in the same locus, and point out that traditional mutation, to some extent, can result in premature convergence of genes (PCG) in the same locus. The above drawback of the traditional mutation operator causes the loss of critical alleles. Inspired by digital technique, we introduce two kinds of boolean operation into GA to develop a novel mutation operator and discuss its contribution to preventing the loss of critical alleles. The experimental results of function optimization show that the improved mutation operator can effectively prevent premature convergence, and can provide a wide selection range of control parameters for GA.
Enhancing Genetic Improvement Mutations Using Large Language Models

Alexander E.I. Brownlee,James Callan,Karine Even-Mendoza,Alina Geiger,Carol Hanna,Justyna Petke,Federica Sarro,Dominik Sobania

DOI: https://doi.org/10.1007/978-3-031-48796-5_13

2023-10-18

Abstract:Large language models (LLMs) have been successfully applied to software engineering tasks, including program repair. However, their application in search-based techniques such as Genetic Improvement (GI) is still largely unexplored. In this paper, we evaluate the use of LLMs as mutation operators for GI to improve the search process. We expand the Gin Java GI toolkit to call OpenAI's API to generate edits for the JCodec tool. We randomly sample the space of edits using 5 different edit types. We find that the number of patches passing unit tests is up to 75% higher with LLM-based edits than with standard Insert edits. Further, we observe that the patches found with LLMs are generally less diverse compared to standard edits. We ran GI with local search to find runtime improvements. Although many improving patches are found by LLM-enhanced GI, the best improving patch was found by standard GI.

Software Engineering,Artificial Intelligence,Machine Learning,Neural and Evolutionary Computing
Towards Objective-Tailored Genetic Improvement Through Large Language Models

Sungmin Kang,Shin Yoo

DOI: https://doi.org/10.48550/arXiv.2304.09386

2023-04-19

Software Engineering

Abstract:While Genetic Improvement (GI) is a useful paradigm to improve functional and nonfunctional aspects of software, existing techniques tended to use the same set of mutation operators for differing objectives, due to the difficulty of writing custom mutation operators. In this work, we suggest that Large Language Models (LLMs) can be used to generate objective-tailored mutants, expanding the possibilities of software optimizations that GI can perform. We further argue that LLMs and the GI process can benefit from the strengths of one another, and present a simple example demonstrating that LLMs can both improve the effectiveness of the GI optimization process, while also benefiting from the evaluation steps of GI. As a result, we believe that the combination of LLMs and GI has the capability to significantly aid developers in optimizing their software.
The Big Mutation:An Operation to Improve the Performance of Genetic Algorithms

Liu Guizhong

1998-01-01

Abstract:With the increasingly wide applications of genetic algorithms (GAs),its instinct drawback,the extraordinarily slow convergence,becomes the biggest obstacle of its further acceptation in many application fields. To combat this problem,a kind of operation,named big mutation,is discussed in this paper. After giving some explanation of the background and implementation of this operation,this paper provides the results of a numerical experiment to prove its efficiency and stability.
Structured mutation inspired by evolutionary theory enriches population performance and diversity

Stefano Tiso,Pedro Carvalho,Nuno Lourenço,Penousal Machado

2023-07-12

Abstract:Grammar-Guided Genetic Programming (GGGP) employs a variety of insights from evolutionary theory to autonomously design solutions for a given task. Recent insights from evolutionary biology can lead to further improvements in GGGP algorithms. In this paper, we apply principles from the theory of Facilitated Variation and knowledge about heterogeneous mutation rates and mutation effects to improve the variation operators. We term this new method of variation Facilitated Mutation (FM). We test FM performance on the evolution of neural network optimizers for image classification, a relevant task in evolutionary computation, with important implications for the field of machine learning. We compare FM and FM combined with crossover (FMX) against a typical mutation regime to assess the benefits of the approach. We find that FMX in particular provides statistical improvements in key metrics, creating a superior optimizer overall (+0.48\% average test accuracy), improving the average quality of solutions (+50\% average population fitness), and discovering more diverse high-quality behaviors (+400 high-quality solutions discovered per run on average). Additionally, FM and FMX can reduce the number of fitness evaluations in an evolutionary run, reducing computational costs in some scenarios.

Neural and Evolutionary Computing
Mathematical Analysis of Mutation Operator and Its Improved Strategy in Genetic Algorithms

Zhang Liangjie,Mao Zhihong,Li Yanda

DOI: https://doi.org/10.1007/s11767-997-1008-2

1997-01-01

Journal of Electronics (China)

Abstract:This paper analyzes the optimization problem of mutation probability in genetic algorithms by applying the definition of i-bit improved sub-space. Then fuzzy reasoning technique is adopted to determine the optimal mutation probability in different conditions. The superior convergence property of the new method is evaluated by applying it to two simulation examples.
Genetic Improvement in the Shackleton Framework for Optimizing LLVM Pass Sequences

Shuyue Stella Li,Hannah Peeler,Andrew N. Sloss,Kenneth N. Reid,Wolfgang Banzhaf

DOI: https://doi.org/10.48550/arXiv.2204.13261

2022-04-28

Neural and Evolutionary Computing

Abstract:Genetic improvement is a search technique that aims to improve a given acceptable solution to a problem. In this paper, we present the novel use of genetic improvement to find problem-specific optimized LLVM pass sequences. We develop a pass-level patch representation in the linear genetic programming framework, Shackleton, to evolve the modifications to be applied to the default optimization pass sequences. Our GI-evolved solution has a mean of 3.7% runtime improvement compared to the -O3 optimization level in the default code generation options which optimizes on runtime. The proposed GI method provides an automatic way to find a problem-specific optimization sequence that improves upon a general solution without any expert domain knowledge. In this paper, we discuss the advantages and limitations of the GI feature in the Shackleton Framework and present our results.
High-Order Quantum Genetic Algorithm Based on Quantum Entanglement

Han Qi,Shuai Gong,Abdullah Gani,Changqing Gong

DOI: https://doi.org/10.1109/eei63073.2024.10695959

2024-01-01

Abstract:Quantum genetic algorithms(QGA) typically construct traditional quantum chromosomes as independent quantum bit systems, which has facilitated the enhancement of search speed and precision in genetic algorithms, demonstrating powerful problem-solving capabilities in optimization issues. Although traditional QGA have enriched population diversity, they still tend to fall into local optima, and the evolutionary update operations require rotation directions to be determined by lookup tables, a process involving multiple conditional judgments that affect algorithm efficiency. To address these issues, an improved method has been proposed to further enhance the performance of QGA. In the initial population selection, a population grouping evolution method based on microhabitat coevolution strategy is used to prevent the algorithm from getting trapped in local optima when solving complex function problems. Additionally, quantum entanglement is introduced on top of quantum super-position by implementing quantum entanglement using quantum registers, and higher-order quantum rotation gates are employed that do not involve multiple conditional judgments of lookup tables, thus improving the efficiency of quantum chromosome measurement and evolutionary updates. Finally, traditional QGA and higher-order quantum genetic algorithms were applied to test functions in multiple experiments. The results indicate that, compared to traditional QGA, the higher-order quantum genetic algorithms have achieved an increase in convergence speed, an approximate 14.3% improvement in convergence precision, and possess enhanced global search capabilities.
A Kind of Genetic Algorithm Based on Compound Mutation Strategy and Performance Study

Fachao Li,Tingyu Zhang,Chenxia Jin

DOI: https://doi.org/10.1109/wicom.2009.5300865

2009-01-01

Abstract:Genetic algorithm (GA), as a kind of important intelligence computing tool, is a wide research content in the academic circle and the application domain now. In this paper, for the mutation operation of GA, by combining with the essential feature, from structural and visualized angle, we implement different mutation strategy to individual of different fitness value in every generation, and establish a genetic algorithm based on compound mutation (denoted by BCM-GA, for short). Further, we discuss the global convergence of BCM-GA by using the Markov chain theory, and analyze the performance of BCM-GA through an example. All the results indicate that, BCM-GA is obviously higher than real number code genetic algorithm (denoted by B10GA, for short) in the convergence time and convergence precision.
An upper bound of the mutation probability in the genetic algorithm for general 0-1 knapsack problem

Yang Yang

2024-03-18

Abstract:As an important part of genetic algorithms (GAs), mutation operators is widely used in evolutionary algorithms to solve $\mathcal{NP}$-hard problems because it can increase the population diversity of individual. Due to limitations in mathematical tools, the mutation probability of the mutation operator is primarily empirically set in practical applications.

Neural and Evolutionary Computing,Computational Complexity,Computer Science and Game Theory
Study on Genetic Algorithm Based on Schema Mutation and Its Performance Analysis

Fachao Li,Tingyu Zhang

DOI: https://doi.org/10.1109/ISECS.2009.129

2009-01-01

Abstract:Genetic algorithm (GA), as a kind of important intelligence computing tool, is a wide research content in the academic circle and the application domain now. In this paper, for the mutation operation of GA, by combining with the essential feature, we establish a genetic algorithm based on schema mutation (denoted by SM-GA, for short). Further, we discuss the global convergence of CM-GA by using the Markov chain theory, and analyze the performance of SM-GA through an example. All the results indicate that, SM-GA is higher than the ordinary binary code genetic algorithm (denoted by B2GA, for short) in convergence precision. There was no significant difference between SM-GA and B2GA in convergence time. SM-GA overcomes the problem that B2GA can not converge strongly to some extent.
Higher-Order Quantum-Inspired Genetic Algorithms

Robert Nowotniak,Jacek Kucharski

DOI: https://doi.org/10.15439/2014f99

2014-09-29

Abstract:This paper presents a theory and an empirical evaluation of Higher-Order Quantum-Inspired Genetic Algorithms. Fundamental notions of the theory have been introduced, and a novel Order-2 Quantum-Inspired Genetic Algorithm (QIGA2) has been developed. Contrary to all QIGA algorithms which represent quantum genes as independent qubits, in higherorder QIGAs quantum registers are used to represent genes strings, which allows modelling of genes relations using quantum phenomena. Performance comparison has been conducted on a benchmark of 20 deceptive combinatorial optimization problems. It has been presented that using higher quantum orders is beneficial for genetic algorithm efficiency, and the new QIGA2 algorithm outperforms the old QIGA algorithm tuned in highly compute-intensive metaoptimization process.
Analysis and comparison of a proposed mutation operator and its effects on the performance of genetic algorithm

Sami Ullah,Abdus Salam,Mohsin Masood

DOI: https://doi.org/10.11591/ijeecs.v25.i2.pp1208-1216

2022-02-01

Indonesian Journal of Electrical Engineering and Computer Science

Abstract:Genetic algorithms (GAs) are dependent on various operators and parameters. The most common evolutionary operators are parent selection, crossover, and mutation. Each operator has broad implementations with its pros and cons. A successful GA is highly dependent on genetic diversity which is the main driving force that steers a GA towards an optimal solution. Mutation operator implements the idea of exploration to search for uncharted areas and introduces diversity in a population. Thus, increasing the probability of GA to converge to a globally optimum solution. In this paper, a new variant of mutation operator is proposed, and its functions are studied and compared with the existing operators. The proposed mutation operator as well as others such as m-mutation, shuffle, swap, and inverse are tested for their ability to introduce diversity in population and hence, their effects on the performance of GA. All these operators are applied to Max one problem. The results concluded that the proposed variant is far more superior to the existing operators in terms of introducing diversity and hence early convergence to an optimum solution.
Mutation-based Test Generation for Quantum Programs with Multi-Objective Search

Xinyi Wang,Tongxuan Yu,Paolo Arcaini,Tao Yue,Shaukat Ali

DOI: https://doi.org/10.1145/3512290.3528869

2022-01-01

Abstract:Mutation testing is often used for designing new tests, and involves changing a program in minor ways, which results in mutated versions of the program, i.e., mutants. An effective test suite should find faults (or kill mutants) with a minimum number of test cases, to save resources required for executing test cases. In this paper, in the context of mutation testing for quantum programs, we present a multi-objective and search-based approach (MutTG) to generate the minimum number of test cases killing as many mutants as possible. MutTG tries to estimate the likelihood that a mutant is equivalent, and uses this as a discount factor in the fitness definition to avoid keeping on trying to kill mutants that cannot be killed. We employed NSGA-II as the multi-objective search algorithm. Then, we compared MutTG with another version of the approach that does not use the discount factor in its fitness definition, and with random search (RS), over a set of open-source quantum programs and their mutants of varying complexity. Results show that the discount factor does indeed help in guiding the test generation, as the approach with the discount factor performs better than the one without it.
Binary-encoding Genetic Algorithms with Modified Mutation Operation

LI Liang-min

DOI: https://doi.org/10.3969/j.issn.1004-731x.2005.05.015

2005-01-01

Abstract:Being a stochastic search technique, there are some shortcomings in standard genetic algorithm such as poor local search ability and premature convergence. An effective method of adapting mutation probabilities on each genome of binary-encoding genetic algorithm is proposed to overcome these shortcomings. At the beginning stage of evolution, larger values of mutation probability are assigned to those genomes in higher bit positions, so that larger feasible region is explored. When the region in which the global optimum exists is found, mutation probabilities of genomes in higher bit positions are decreased to prevent better solutions from disruption. Accordingly, mutation probabilities of genomes in lower bit positions are increased to improve the local search ability. Experiments results on five multimodal functions demonstrate that proposed genetic algorithm possesses better optimization ability and gets stuck at a local optimum fewer times than standard genetic algorithm does.
A theoretical analysis on mutation operator of standard genetic algorithm

Guo Li,Zongben Xu,l p guo

DOI: https://doi.org/10.1007/BF02900893

1996-01-01

Wuhan University Journal of Natural Sciences

Abstract:A theoretical analysis on the function of the mutation operator to GAs is developed, and the probability in which GAs converge to the premature set is also estimated. Our discussion indícates that mutation is necessary to GAs.
Improved Genetic Algorithms' and Its Application in MATLAB

Ying LI,Qi WANG,Yuan-bo WANG,Bi-bo RAO

DOI: https://doi.org/10.3969/j.issn.1002-2333.2011.11.018

2011-01-01

Abstract:Aiming at appearing premature and fluctuations in the traditional genetic algorithms(GA),three genetic manipulations of influence GA's accuracy and efficiency selection,crossover and mutation are analyzed.By using SUS,three crossovers and MUT,the convergence accuracy and speed are improved.The feasibility of its improvement is tested by using MATLAB,and the effectiveness of improved genetic algorithms is verified.
A Novel and Accelerated Genetic Algorithm

Huang Bao-Juan,Zhuang Jian,Yu De-Hong

2008-01-01

WSEAS TRANSACTIONS ON SYSTEMS AND CONTROL

Abstract:Genetic algorithm (GA) is very helpful when the developer does not have precise domain expertise, because GA possesses the ability to explore and learn from their domain. At present, the research of GA mainly focuses on the three operators and devotes to improve the algorithm efficiency and avoid premature convergence. This paper presents a cycle mutation operator and a novel selection operator; accordingly, an improved cycle mutation genetic algorithm (ICMGA) is schemed, The experimental results compared with other genetic algorithms validate the performance of this algorithm, such as the exploration ability in search space, the stabilization and calculation speed, are all superior to other algorithms, and ICMGA is not sensitive to the initial population distribution.
Semantically-Oriented Mutation Operator in Cartesian Genetic Programming for Evolutionary Circuit Design

David Hodan,Vojtech Mrazek,Zdenek Vasicek

DOI: https://doi.org/10.1145/3377930.3390188

2020-04-23

Abstract:Despite many successful applications, Cartesian Genetic Programming (CGP) suffers from limited scalability, especially when used for evolutionary circuit design. Considering the multiplier design problem, for example, the 5x5-bit multiplier represents the most complex circuit evolved from a randomly generated initial population. The efficiency of CGP highly depends on the performance of the point mutation operator, however, this operator is purely stochastic. This contrasts with the recent developments in Genetic Programming (GP), where advanced informed approaches such as semantic-aware operators are incorporated to improve the search space exploration capability of GP. In this paper, we propose a semantically-oriented mutation operator (SOMO) suitable for the evolutionary design of combinational circuits. SOMO uses semantics to determine the best value for each mutated gene. Compared to the common CGP and its variants as well as the recent versions of Semantic GP, the proposed method converges on common Boolean benchmarks substantially faster while keeping the phenotype size relatively small. The successfully evolved instances presented in this paper include 10-bit parity, 10+10-bit adder and 5x5-bit multiplier. The most complex circuits were evolved in less than one hour with a single-thread implementation running on a common CPU.

Neural and Evolutionary Computing

Genetic Improvement using Higher Order Mutation

Parameter estimation for time-varying system based on improved genetic algorithm

Improved Genetic Operator for Genetic Algorithm

Enhancing Genetic Improvement Mutations Using Large Language Models

Towards Objective-Tailored Genetic Improvement Through Large Language Models

The Big Mutation:An Operation to Improve the Performance of Genetic Algorithms

Structured mutation inspired by evolutionary theory enriches population performance and diversity

Mathematical Analysis of Mutation Operator and Its Improved Strategy in Genetic Algorithms

Genetic Improvement in the Shackleton Framework for Optimizing LLVM Pass Sequences

High-Order Quantum Genetic Algorithm Based on Quantum Entanglement

A Kind of Genetic Algorithm Based on Compound Mutation Strategy and Performance Study

An upper bound of the mutation probability in the genetic algorithm for general 0-1 knapsack problem

Study on Genetic Algorithm Based on Schema Mutation and Its Performance Analysis

Higher-Order Quantum-Inspired Genetic Algorithms

Analysis and comparison of a proposed mutation operator and its effects on the performance of genetic algorithm

Mutation-based Test Generation for Quantum Programs with Multi-Objective Search

Binary-encoding Genetic Algorithms with Modified Mutation Operation

A theoretical analysis on mutation operator of standard genetic algorithm

Improved Genetic Algorithms' and Its Application in MATLAB

A Novel and Accelerated Genetic Algorithm

Semantically-Oriented Mutation Operator in Cartesian Genetic Programming for Evolutionary Circuit Design