## Tackling Real-Coded Genetic Algorithms: Operators and Tools for Behavioural Analysis (1998)

Venue: | Artificial Intelligence Review |

Citations: | 128 - 24 self |

### BibTeX

@ARTICLE{Herrera98tacklingreal-coded,

author = {F. Herrera and M. Lozano and J. L. Verdegay},

title = {Tackling Real-Coded Genetic Algorithms: Operators and Tools for Behavioural Analysis},

journal = {Artificial Intelligence Review},

year = {1998},

volume = {12},

pages = {265--319}

}

### Years of Citing Articles

### OpenURL

### Abstract

. Genetic algorithms play a significant role, as search techniques for handling complex spaces, in many fields such as artificial intelligence, engineering, robotic, etc. Genetic algorithms are based on the underlying genetic process in biological organisms and on the natural evolution principles of populations. These algorithms process a population of chromosomes, which represent search space solutions, with three operations: selection, crossover and mutation. Under its initial formulation, the search space solutions are coded using the binary alphabet. However, the good properties related with these algorithms do not stem from the use of this alphabet; other coding types have been considered for the representation issue, such as real coding, which would seem particularly natural when tackling optimization problems of parameters with variables in continuous domains. In this paper we review the features of real-coded genetic algorithms. Different models of genetic operators and some me...

### Citations

7411 |
Genetic Algorithms
- Goldberg
- 1989
(Show Context)
Citation Context ...c algorithm 1. Introduction Genetic algorithms (GAs) are general purpose search algorithms which use principles inspired by natural genetic populations to evolve solutions to problems (Holland, 1975; =-=Goldberg, 1989-=-a). The basic idea is to maintain a population of chromosomes, which represent candidate solutions to the concrete problem, that evolves over time through a process of competition and controlled varia... |

1979 |
Genetic Algorithms + Data Structure = Evolution Programs, 3 ed
- Michalewicz
- 1996
(Show Context)
Citation Context ...f metaoperators in order to find the most adequate parameters for a standard GA. Subsequently, RCGAs have been mainly used for numerical optimization on continuous domains (Wright, 1991; Davis, 1991; =-=Michalewicz, 1992-=-; Eshelman et al., 1993; Muhlenbein et al., 1993; Herrera et al., 1994; Herrera et al., 1995). Until 1991 specific theoretical studies about RCGA operation weren't done and so the use of these algorit... |

1178 | Handbook of Genetic Algorithms
- Davis
- 1991
(Show Context)
Citation Context ...for the use of metaoperators in order to find the most adequate parameters for a standard GA. Subsequently, RCGAs have been mainly used for numerical optimization on continuous domains (Wright, 1991; =-=Davis, 1991-=-; Michalewicz, 1992; Eshelman et al., 1993; Muhlenbein et al., 1993; Herrera et al., 1994; Herrera et al., 1995). Until 1991 specific theoretical studies about RCGA operation weren't done and so the u... |

872 | Genetic Programming
- Koza
- 1992
(Show Context)
Citation Context ... a key issue in GA work because GAs directly manipulate a coded representation of the problem and because the representation schema can severely limit the window by which a system observes its world (=-=Koza, 1992-=-). Fixed-length and binary coded strings for the representation solution have dominated GA research since there are theoretical results that show them to be the most effective ones (Goldberg, 1991a), ... |

870 | An analysis of the behavior of a class of genetic adaptive systems. Doctoral dissertation - Jong - 1975 |

530 |
Handbook of Genetic Algorithms
- Syswerda
- 1991
(Show Context)
Citation Context ...eneralization of the simple crossover; n crossover points are randomly selected and the segments of the parents, defined by them, are exchanged for generating the offspring. \Gamma uniform crossover (=-=Syswerda, 1989-=-): the values of each gene in the offspring are determined by the uniform random choice of the values of this gene in the parents. Other types of crossover operators under binary coding are reported i... |

395 | A comparative analysis of selection schemes used in genetic algorithms - Goldberg, Deb - 1991 |

344 | D.: Predictive models for the breeder genetic algorithm I. continuous parameter optimization - Mühlenbein, Schlierkamp-Voosen - 1993 |

229 | H.P.: A survey of evolution strategies
- Back, Hoffmeister, et al.
- 1991
(Show Context)
Citation Context ...sociated with C i . Figure 3 shows the roulette wheel associated with the selection process of Figure 2 when stochastic universal sampling is used. A wide set of selection mechanisms are reviewed in (=-=Back et al., 1991-=-b). air96.tex; 14/03/1996; 19:57; no v.; p.6 REAL-CODED GAs 7 2.4. Recombination Through Crossover and Mutation After selection has been carried out, the construction of the intermediate population is... |

212 |
Adaptive selection methods for genetic algorithms
- Baker
- 1985
(Show Context)
Citation Context ...C i ) = f(C i ) P N j=1 f(C j ) : In this way, chromosomes with above-average fitness tend to receive more copies than those with below-average fitness. A different approach is the ranking selection (=-=Baker, 1985-=-); the chromosomes are sorted in order of raw fitness, and then p s (C i ), i = 1; :::; N , is computed according to the rank of C i by using a non-increasing assignment function. 2.3.2. Sampling Algo... |

212 | Global Optimization - Torn, Zilinskas - 1989 |

197 |
Classifier Systems and Genetic Algorithms
- Booker, Goldberg, et al.
- 1989
(Show Context)
Citation Context ...(Goldberg, 1989a; Davis, 1991), numerical function and combinatorial optimization (Goldberg, 1989a; Michalewicz, 1992; Davis, 1991), robotic (Davidor, 1991), classifier systems (Holland et al., 1986; =-=Booker et al., 1989-=-; Belew et al., 1991; Forrest, 1993), learning (Grefenstette, 1990), pattern recognition (Forrest et al., 1993), neuronal networks (Whitley et al., 1992; Forrest, 1993), fuzzy systems (Cordon et al., ... |

184 |
Adapting operator probabilities in genetic algorithms
- Davis
- 1989
(Show Context)
Citation Context ...n real number representation are called real-coded GAs (RCGAs). The use of real coding initially appears in specific applications, such as in (Lucasius et al., 1989) for chemometric problems, and in (=-=Davis, 1989-=-) for the use of metaoperators in order to find the most adequate parameters for a standard GA. Subsequently, RCGAs have been mainly used for numerical optimization on continuous domains (Wright, 1991... |

183 |
An introduction to simulated evolutionary optimization
- Fogel
- 1994
(Show Context)
Citation Context ... that a single gene may simultaneously affect several phenotype traits. Polygeny is the effect when a single phenotypic characteristic may be determined by the simultaneous interaction of many genes (=-=Fogel, 1994-=-). The model in (Voigt, 1992) and (Voigt, 1993) is an attempt to include this last effect in the GA's behaviour. During the 1960s, a research line of algorithms based on natural evolution, called Evol... |

153 |
Genetic algorithms and Walsh functions: Part 1, A gentle introduction
- Goldberg
- 1989
(Show Context)
Citation Context ...s under binary coding are reported in (Eshelman et al., 1989). 3.3. Mutation Operator Given a chromosome, a gene is randomly chosen and its value is swapped; "1" for "0" and viceve=-=rsa (Holland, 1975; Goldberg, 1989-=-a). Figure 5 shows an example of the application of mutation operators. 3.4. Example of BCGA Iteration Next, we show an example of an iteration of a BCGA during only one generation for the function f(... |

139 | Real-coded genetic algorithms and intervalschemata - Eshelman, Schaffer |

133 | Genetic algorithms for real parameter optimization," Proceedings of Foundations of Genetic Algorithms,(edited by Gregory
- Wright
- 1991
(Show Context)
Citation Context ...993) h i = c 1 i + ff i (c 2 i \Gamma c 1 i ) and ff i is a randomly (uniformly) chosen value in the interval [\Gamma0:25; 1:25]. This operator is equal to the BLX-0.25. Wright's heuristic crossover (=-=Wright, 1990-=-) Let's suppose that C 1 is the parent with the best fitness. Then h i = r \Delta (c 1 i \Gamma c 2 i ) + c 1 i and r is a random number belonging to [0; 1]. Linear BGA crossover (Schlierkamp-Voosen, ... |

102 | Equivalence class analysis of genetic algorithms
- Radcliffe
(Show Context)
Citation Context ...nce this suggested that binary coding should be more effective than codings based on large alphabets. Later, tools for the theoretical treatment of RCGAs were proposed (Wright, 1991; Goldberg, 1991a; =-=Radcliffe, 1991-=-a; Eshelman et al., 1993) and so their power was corroborated. The main objective of this paper is to deal with the RCGAs. To do that, first we study the binary coding and its advantages and drawbacks... |

97 |
Biases in the crossover landscape
- Eshelman, Caruna, et al.
- 1989
(Show Context)
Citation Context ...ssover Figure 5. Mutation Numerous research papers have been directed at finding alternative more powerful crossover operators for BCGAs. Two of the most important ones are: \Gamma n-point crossover (=-=Eshelman et al., 1989-=-): this is a generalization of the simple crossover; n crossover points are randomly selected and the segments of the parents, defined by them, are exchanged for generating the offspring. \Gamma unifo... |

86 | Dynamic Parameter Encoding for Genetic Algorithms - Schraudolph, Belew - 1992 |

78 | Real-coded Genetic Algorithms, Virtual Alphabets
- Goldberg
- 1991
(Show Context)
Citation Context ...search techniques. Fixed-length and binary coded strings for the representation solution have dominated GA research since there are theoretical results that show them to be the most appropriate ones (=-=Goldberg, 1991-=-a), and as they are amenable to simple implementation. But the GA's good properties do not stem from the use of bit strings (Antonisse, 1989; Radcliffe, 1992). For this reason, the path has been lain ... |

77 |
Genetic and evolutionary algorithms come of age
- Goldberg
- 1994
(Show Context)
Citation Context ...search techniques. Fixed-length and binary coded strings for the representation solution have dominated GA research since there are theoretical results that show them to be the most appropriate ones (=-=Goldberg, 1991-=-a), and as they are amenable to simple implementation. But the GA's good properties do not stem from the use of bit strings (Antonisse, 1989; Radcliffe, 1992). For this reason, the path has been lain ... |

71 |
A new interpretation of schema notation that overturns the binary encoding constraint
- Antonisse
- 1989
(Show Context)
Citation Context ...ical results that show them to be the most appropriate ones (Goldberg, 1991a), and as they are amenable to simple implementation. But the GA's good properties do not stem from the use of bit strings (=-=Antonisse, 1989-=-; Radcliffe, 1992). For this reason, the path has been lain toward the use of non-binary representations more adequate for each particular application problem. One of the most important ones is the re... |

59 |
Tuning fuzzy logic controllers by genetic algorithms
- Herrera, Lozano, et al.
- 1995
(Show Context)
Citation Context ...ntly, RCGAs have been mainly used for numerical optimization on continuous domains (Wright, 1991; Davis, 1991; Michalewicz, 1992; Eshelman et al., 1993; Muhlenbein et al., 1993; Herrera et al., 1994; =-=Herrera et al., 1995-=-). Until 1991 specific theoretical studies about RCGA operation weren't done and so the use of these algorithms was controversial; researchers familiar with fundamental GA theory didn't understand the... |

58 | Representational issues in genetic optimization - Liepins, Vose - 1990 |

57 | An Experimental Comparison of Binary and Floating Point Representations in Genetic Algorithms - Janikow, Michalewicz - 1991 |

55 |
D.: Scheduling problems and traveling salesmen: The genetic edge recombination operator
- Whitley, Starkweather, et al.
- 1989
(Show Context)
Citation Context ...(Bramlette et al., 1991), unsupervised learning of neural networks (Ichikawa et al., 1993), etc. \Gamma ordered lists, for schedule optimization problems (Syswerda, 1991), traveling salesman problem (=-=Whitley et al., 1989-=-), job-shop scheduling problems (Fox et al., 1991) \Gamma lisp expressions, for evolving computer programs for control tasks, robotic planning and symbolic regression (Koza, 1992). \Gamma two-dimentio... |

53 | An overview of genetic algorithms: Part 2, research topics
- Beasley, Bull, et al.
- 1993
(Show Context)
Citation Context ... not be guaranteed that after the application of the crossover and mutation operators the codes generated are not redundant. Mechanisms to limit these types of codes in the population are needed. In (=-=Beasley et al., 1993-=-) the following possibilities are reviewed: 1) discard the redunair96. tex; 14/03/1996; 19:57; no v.; p.18 REAL-CODED GAs 19 dant codes generated or 2) assign low fitness to the chromosomes with redun... |

53 | Messy Genetic Algorithms
- Goldberg, Korb, et al.
- 1989
(Show Context)
Citation Context ...ve. It is due since the schemata that have the global optimum don't bear significance and so they may not proliferate during the genetic process. Example 5 shows a problem where deception is present (=-=Goldberg et al., 1989-=-). EXAMPLE 5. Let us consider the problem shown in Table V. For this problem the following relationships are hold: f(0 ? ?) ? f(1 ? ?) f(00?) ? f(11?), f(01?), f(10?) f(?0?) ? f(?1?) f(0 ? 0) ? f(1 ? ... |

45 | Genetic Operators for Sequencing Problems - Fox, McMahon - 1991 |

44 | Extended selection mechanisms in genetic algorithms
- Back, Hoffmeister
- 1991
(Show Context)
Citation Context ...sociated with C i . Figure 3 shows the roulette wheel associated with the selection process of Figure 2 when stochastic universal sampling is used. A wide set of selection mechanisms are reviewed in (=-=Back et al., 1991-=-b). air96.tex; 14/03/1996; 19:57; no v.; p.6 REAL-CODED GAs 7 2.4. Recombination Through Crossover and Mutation After selection has been carried out, the construction of the intermediate population is... |

42 |
Genetic Algorithms And Robotics - A Heuristic Strategy For Optimization. World Scientific Publishing Co. Pte
- Davidor
- 1991
(Show Context)
Citation Context ...cations have enormously grown in many fields: engineering (Goldberg, 1989a; Davis, 1991), numerical function and combinatorial optimization (Goldberg, 1989a; Michalewicz, 1992; Davis, 1991), robotic (=-=Davidor, 1991-=-), classifier systems (Holland et al., 1986; Booker et al., 1989; Belew et al., 1991; Forrest, 1993), learning (Grefenstette, 1990), pattern recognition (Forrest et al., 1993), neuronal networks (Whit... |

37 |
The ARGOT strategy: Adaptive representation genetic optimizer technique
- Shaefer
- 1987
(Show Context)
Citation Context ...timum. Since during each run the action interval of the parameters is limited, the precision shall be more refined. Two techniques based on this idea for refining precision have been proposed: ARGOT (=-=Shaefer, 1987-=-), dynamic parameter encoding (Schraudolph et al., 1992) and delta coding (Whitley et al., 1991). An effect that appears using the binary alphabet for representing elements in continuous domains is th... |

34 | Hybridizing genetic algorithms with hill-climbing methods for global optimization: Two possible ways - Renders, Bersini - 1994 |

32 | Delta Coding: An iterative search strategy for genetic algorithms
- Whitley, Mathias, et al.
- 1991
(Show Context)
Citation Context ...sion shall be more refined. Two techniques based on this idea for refining precision have been proposed: ARGOT (Shaefer, 1987), dynamic parameter encoding (Schraudolph et al., 1992) and delta coding (=-=Whitley et al., 1991-=-). An effect that appears using the binary alphabet for representing elements in continuous domains is the so-called Hamming cliff. It is produced when the binary coding of two adjacent values differs... |

29 |
Schedule optimization using genetic algorithms Handbook of Genetic Algorithms
- Syswerda
- 1991
(Show Context)
Citation Context ...mlette, 1991), parametric design of aircraft (Bramlette et al., 1991), unsupervised learning of neural networks (Ichikawa et al., 1993), etc. \Gamma ordered lists, for schedule optimization problems (=-=Syswerda, 1991-=-), traveling salesman problem (Whitley et al., 1989), job-shop scheduling problems (Fox et al., 1991) \Gamma lisp expressions, for evolving computer programs for control tasks, robotic planning and sy... |

28 | Theoretical analysis of evolutionary algorithms with infinite population size in continuous space, part II: Analysis of the diversification role of crossover - Xiaofang, Palmieri - 1994 |

24 |
A User’s Guide to GENESIS Version 5.0
- Grefenstette
- 1990
(Show Context)
Citation Context ...s used by each one of FCB crossovers considered in Table VIII. A binary-coded GA (BCGA) has also been included, which is based on a two point crossover. For this purpose, we used the GENESIS program (=-=Grefenstette, 1990-=-). The number of binary genes assigned to each variable is 11 for f 1 and f 2 and 31 for f 3 Therefore, the precision, ae, is approximately 10 \Gamma6 . air96.tex; 14/03/1996; 19:57; no v.; p.31 32 F.... |

24 |
Pictorial Representations of fuzzy connectives, part II
- Mizumoto
- 1989
(Show Context)
Citation Context ...tor to them. We may generate the chromosome H = (h 1 ; :::; h i ; :::; h n ) as h i = Q(c 1 i ; c 2 i ); i = 1; :::; n: T-norms, t-conorms, averaging functions and generalized compensation operators (=-=Mizumoto, 1989-=-a; Mizumoto, 1989b) may be used to obtain F , S, M and L operators. We shall associate F to a t-norm T , S to a t-conorm G, M with an averaging operator P , and L with a generalized compensation opera... |

23 |
F.(1991) ‘Initialisation, mutation and selection methods in Genetic algorithms for function optimisation
- Bramlette
(Show Context)
Citation Context ...oran et al., 1994), tuning fuzzy logic controllers (Herrera et al., 1995), etc. air96.tex; 14/03/1996; 19:57; no v.; p.4 REAL-CODED GAs 5 \Gamma vectors of integer numbers, for function optimization (=-=Bramlette, 1991-=-), parametric design of aircraft (Bramlette et al., 1991), unsupervised learning of neural networks (Ichikawa et al., 1993), etc. \Gamma ordered lists, for schedule optimization problems (Syswerda, 19... |

21 |
A Genetic Algorithm for the Linear Transportation Problem
- Vignaux, Michalewicz
- 1991
(Show Context)
Citation Context ...ions, for evolving computer programs for control tasks, robotic planning and symbolic regression (Koza, 1992). \Gamma two-dimentional matrix of integer numbers, for the linear transportation problem (=-=Michalewicz, 1991-=-). 2.3. Selection Mechanism Let's consider P a population with chromosomes C 1 ; :::; CN . The selection mechanism produces a intermediate population, P 0 , with copies of chromosomes in P . The numbe... |

20 | Expected allele coverage and role of mutation in genetic algorithms - Tate, Smith - 1993 |

16 | editor. Genetic Algorithms for Machine Learning - Grefenstette - 1995 |

16 |
Fuzzy evolutionary algorithms
- Voigt
- 1992
(Show Context)
Citation Context ... the variables which they represent, so the genetic operators must observe this requirement. There are some representation models which are discarded from the previous one, one gene-one variable. In (=-=Voigt, 1992-=-) and (Voigt, 1993) a coding type is described under which each problem parameter has associated a number of m decision genes belonging to the interval [0; 1]. The chromosomes are formed by the link o... |

14 | Fuzzy Tools to Improve Genetic Algorithms
- Herrera, Lozano, et al.
- 1994
(Show Context)
Citation Context ... standard GA. Subsequently, RCGAs have been mainly used for numerical optimization on continuous domains (Wright, 1991; Davis, 1991; Michalewicz, 1992; Eshelman et al., 1993; Muhlenbein et al., 1993; =-=Herrera et al., 1994-=-; Herrera et al., 1995). Until 1991 specific theoretical studies about RCGA operation weren't done and so the use of these algorithms was controversial; researchers familiar with fundamental GA theory... |

14 | Characterizing crossover in genetic algorithms - Liepins, Vose - 1992 |

12 | Genetic Algorithms in Parametric Design of Aircraft - Bramlette, Bouchard |

12 |
Using Genetic Algorithms to Explore
- Forrest, Javornik, et al.
- 1993
(Show Context)
Citation Context ...cz, 1992; Davis, 1991), robotic (Davidor, 1991), classifier systems (Holland et al., 1986; Booker et al., 1989; Belew et al., 1991; Forrest, 1993), learning (Grefenstette, 1990), pattern recognition (=-=Forrest et al., 1993-=-), neuronal networks (Whitley et al., 1992; Forrest, 1993), fuzzy systems (Cordon et al., 1995), artificial life (Belew et al., 1991), etc. 3. Binary-Coded Genetic Algorithms This section is devoted t... |

8 | Aplications of genetic algorithm in chemometrics - LUCASIUS, KATEMAN - 1989 |

7 | Representation and hidden bias: Gray versus binary codign for genetic algorithms - Caruana, Schaffer - 1988 |