D.E.Goldberg (1989) Genetic algorithms and Walsh functions: part II, deception and its analysis. Complex Systems, 3, 153-171.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....to show how this viewpoint enables us to gain further insights into the determination of epistatic effects, and into the value of different forms of encoding a problem for a GA solution. We also demonstrate the equivalence of this approach to the Walsh transform analysis popularized by Goldberg [3, 4], and its extension to the idea of partition coefficients [5] We then show how the experimental design perspective helps to throw further light on the nature of deception. 1 INTRODUCTION The term epistasis is used in the field of genetic algorithms to denote the effect on chromosome fitness of ....

....degree of non linearity in the fitness function, and roughly speaking, the more epistatic the problem is, the harder it may be for a GA to find its optimum. Table 1: Goldberg s 3 bit deceptive function String Fitness 0 0 0 7 0 0 1 5 0 1 0 5 0 1 1 0 1 0 0 3 1 0 1 0 1 1 0 0 1 1 1 8 Several authors [3, 4, 6, 8] have explored the problem of epistasis in terms of the properties of a particular class of epistatic problems, those known as deceptive problems the most famous example of which is probably Goldberg s 3 bit function, which has the form shown in Table 1 (definitions of this function in the ....

[Article contains additional citation context not shown here]

D.E.Goldberg (1989) Genetic algorithms and Walsh functions: part II, deception and its analysis. Complex Systems, 3, 153-171.

....of collateral noise in GA population [15] and the population sizing equation [5] are some instances that directly relate themselves with the signal to noise framework. An earlier effort [15] also resulted in interesting formulation of signal and noise in GA based on static walsh analysis [4]. Although the analysis was restricted, it provided important insights and motivated me to pursue a more generalized and unified framework to address the role of signal and noise in GA. In this paper we present a general perspective of signal and noise from different partitions, which may be ....

....Apart from isolation, deception also introduces a potential source of trouble by making sure that first order estimates of shorter schema distribution is not enough to identify a class to be better than some other class. Deceptive problems are also widely reported to be difficult to solve using GA [4] [12] Isolation of desired solutions result in an initial signal detection problem. Therefore, unless the initial population is large enough to make a correct decision at the order, above the order of deception, the signal is bound to be in the wrong direction. Figure 5 shows the signal to noise ....

D. E. Goldberg. Genetic algorithms and Walsh functions: Part II, deception and its analysis, Complex Systems, 3, 129-152, 1989.

....a universe in which both antigens and antibodies (more precisely, receptors on B cells and T cells) are represented by binary strings. The model uses genetic algorithms (GAs) an idealized computational model of Darwinian evolution based on the principles of genetic variation and natural selection [16, 10]. The GA without crossover is a reasonable model of clonal selection, while the GA with crossover models genetic evolution. In this paper we formulate abstract versions of pattern recognition problems that the immune system appears capable of solving. We then solve these problems using the GA. Our ....

D. E. Goldberg. Genetic algorithms and walsh functions: Part ii, deception and its analysis. Complex Systems, 3:153--171, 1989.

....and attempted to get the linkage right prior to subsequent genetic processing. Those efforts were fairly successful, apparently achieving global solutions in polynomial times in a sense similar to that of the probably almost correct (PAC) algorithms of computational learning theory (Deb, 1991; Goldberg, Deb, Korb, 1990, 1991) In some sense these efforts were both good news and bad news. The good news was that hard problems could be solved in polynomial time, and this offered promise that perhaps all problems of bounded difficulty could be solved as quickly. The bad news was that there appeared to be a mismatch ....

....the investigation necessary to determine whether these results carry over to arbitrary problems of bounded difficulty. 2 A Brief Review of Messy GAs In this section, we briefly review messy GAs. Readers interested in more detail should consult other sources (Deb, 1991; Deb Goldberg, 1991; Goldberg, Deb, Korb, 1990, 1991; Goldberg Kerzic, 1990; Goldberg, Korb, Deb, 1989) Specifically, the following topics are reviewed: 1. messy codes; 2. handling over and underspecification; 3. mGA inner and outer loops; 4. basic mGA theory; 5. time complexity estimates. In the remainder of this section, each of these ....

[Article contains additional citation context not shown here]

Goldberg, D. E. (1989b). Genetic algorithms and Walsh functions: Part II, deception and its analysis.

No context found.

Goldberg, D. E. (1989c). Genetic algorithms and Walsh functions: Part II, deception and its analysis. Complex Systems, 3, 129-152.

No context found.

D. E. Goldberg (1989c). Genetic algorithms and Walsh functions: Part II, Deception and its analysis.

Documents 51 to 61  Previous 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC