## Fitness Distance Correlation as a Measure of Problem Difficulty for Genetic Algorithms (1995)

Venue: | Proceedings of the Sixth International Conference on Genetic Algorithms |

Citations: | 215 - 4 self |

### BibTeX

@INPROCEEDINGS{Jones95fitnessdistance,

author = {Terry Jones and Stephanie Forrest},

title = {Fitness Distance Correlation as a Measure of Problem Difficulty for Genetic Algorithms},

booktitle = {Proceedings of the Sixth International Conference on Genetic Algorithms},

year = {1995},

pages = {184--192},

publisher = {Morgan Kaufmann}

}

### Years of Citing Articles

### OpenURL

### Abstract

A measure of search difficulty, fitness distance correlation (FDC), is introduced and examined in relation to genetic algorithm (GA) performance. In many cases, this correlation can be used to predict the performance of a GA on problems with known global maxima. It correctly classifies easy deceptive problems as easy and difficult non-deceptive problems as difficult, indicates when Gray coding will prove better than binary coding, and is consistent with the surprises encountered when GAs were used on the Tanese and royal road functions. The FDC measure is a consequence of an investigation into the connection between GAs and heuristic search. 1 INTRODUCTION A correspondence between evolutionary algorithms and heuristic state space search is developed in (Jones, 1995b). This is based on a model of fitness landscapes as directed, labeled graphs that are closely related to the state spaces employed in heuristic search. We examine one aspect of this correspondence, the relationship between...

### Citations

8314 | Genetic Algorithms - Goldberg - 1989 |

1215 | The origins of order: self-organization and selection in evolution - Kauffman - 1993 |

1077 | A formal basis for the heuristic determination of minimum cost paths - Hart, Nilsson, et al. - 1968 |

927 | An Analysis of the Behavior of a Class of Genetic Adaptive Systems - Jong - 1975 |

728 |
Heuristics: Intelligent Search Strategies for Computer Problem Solving
- Pearl
- 1984
(Show Context)
Citation Context ...n. In GAs, this function is a fitness function and in heuristic search it is an heuristic function. In heuristic statespace search, there is a large body of work on properties of heuristic functions (=-=Pearl, 1984-=-). A general principle of heuristic functions is that they should correlate well with the distance to the goal of the search, as was suggested as early as 1966 by Doran and Michie (1966). Heuristic se... |

338 | Messy Genetic Algorithms: Motivation, Analysis and First Results - Goldberg, Korb, et al. - 1989 |

254 | A Connectionist Machine for Genetic Hillclimbers - Ackley - 1987 |

210 | The royal road for genetic algorithms: Fitness landscapes and GA performance - Mitchell, Forrest, et al. - 1992 |

173 |
Evolutionary Algorithms, Fitness Landscapes and Search
- Jones
- 1995
(Show Context)
Citation Context ...consequence of an investigation into the connection between GAs and heuristic search. 1 INTRODUCTION A correspondence between evolutionary algorithms and heuristic state space search is developed in (=-=Jones, 1995-=-b). This is based on a model of fitness landscapes as directed, labeled graphs that are closely related to the state spaces employed in heuristic search. We examine one aspect of this correspondence, ... |

148 | Correlated and uncorrelated fitness landscapes and how to tell the difference - Weinberger - 1990 |

129 | Simple genetic algorithm and the minimal deceptive problem - Goldberg - 1987 |

119 | A new adaptive multi-start technique for combinatorial global optimizations - Boese, Kahng, et al. - 1994 |

118 | The genetic algorithm and the structure of the fitness landscape - Manderick, Weger, et al. - 1991 |

110 | What Makes a Problem Hard for a Genetic Algorithm? Some Anomalous Results and Their Explanation Machine Learning 13 - Forrest, Mitchell - 1993 |

101 | Epistasis Variance: A Viewpoint on GA-Hardness", in Foundations of Genetic Algorithms - Davidor - 1991 |

64 | Genetic Algorithm as Function Optimizers - Bethke - 1980 |

62 | On non-computable functions - Rado - 1962 |

61 | Representation and hidden bias: Gray vs. binary coding for genetic algorithms - Caruana, Schaffer - 1988 |

60 | Genetic algorithm difficulty and the modality of fitness landscapes, FOGA-3 - Horn, Goldberg - 1994 |

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

58 | Distributed Genetic Algorithm for Function Optimization - Tanese - 1989 |

57 | Sufficient conditions for deceptive and easy binary functions - Deb, Goldberg - 1993 |

55 | representational bias, and test suite design - Davis - 1991 |

42 | Recombination, Selection, and the Genetic Construction of Computer Programs - Tackett - 1994 |

41 | Cost Versus Distance in the Traveling Salesman Problem - Boese - 1995 |

36 | Experiments with the Graph Traverser Program - Doran, Michie - 1966 |

36 | Genetic Algorithms - Forrest - 1996 |

33 |
A description of hollandâ€™s royal road function
- Jones
- 1995
(Show Context)
Citation Context ...consequence of an investigation into the connection between GAs and heuristic search. 1 INTRODUCTION A correspondence between evolutionary algorithms and heuristic state space search is developed in (=-=Jones, 1995-=-b). This is based on a model of fitness landscapes as directed, labeled graphs that are closely related to the state spaces employed in heuristic search. We examine one aspect of this correspondence, ... |

27 |
The only challenging problems are deceptive: Global search by solving order-1 hyperplanes
- Das, Whitley
- 1991
(Show Context)
Citation Context ...ption to GAs is a contentious issue. Conflicting and extreme statements have been made, ranging from claims that deception is the only thing that is important in making a problem hard for a GA (Das & =-=Whitley, 1991-=-), through claims that deception is neither necessary nor sufficient for a problem to be hard for a GA (Grefenstette, 1993), to informal claims that deception is irrelevant to real-world problems. It ... |

21 | Spurious correlations and premature convergence in genetic algorithms - Schaffer, Eshelman - 1991 |

11 | Signal-to-noise, Crosstalk and Long Range Problem Difficulty in Genetic Algorithms - Kargupta - 1995 |

10 | Towards a stronger building-blocks hypothesis: Effects of relative building-block fitness - Forrest, Mitchell - 1993 |

1 |
Deception Considered Harmful. Pages 75--91 of
- Grefenstette
- 1993
(Show Context)
Citation Context ...ption is the only thing that is important in making a problem hard for a GA (Das & Whitley, 1991), through claims that deception is neither necessary nor sufficient for a problem to be hard for a GA (=-=Grefenstette, 1993-=-), to informal claims that deception is irrelevant to real-world problems. It is clear that some deceptive problems are hard but also that there are other factors that cause difficulty for a GA, such ... |

1 | Aug 12). Royal Road Functions. Internet Genetic Algorithms Digest v7n22. Available via ftp from ftp.santafe.edu in pub/terry/jhrr.tar.gz - Holland - 1993 |

1 | Long Path Problems. Pages 149--158 of - Horn, Goldberg - 1994 |

1 | Adaptation on Rugged Fitness Landscapes. Pages 527--618 - Kauffman - 1989 |

1 | Deceptiveness and Genetic Algorithm Dynamics. Pages 36--50 of - Vose - 1991 |