T Cell–Inspired Algorithms - Immunological Computation: Theory and Applications - page 97

25 else

26 go to 31

27 end if

28 end if

29 end for

30 C ← C ∪ < x, r > {save a new candidate}

=

31 if N n then

← ∪

32 D D C

33 go to 4

34 else

35 go to 7

36 end if
37 until |D| = Tmax {Exception case: too many detectors to handle}

4.6.6 Multishaped Negative Detector Generation

h is algorithm extends the model of RNS by incorporating multiple hypershape
(hypersphere, hyperrectangle, or hyperellipse) detector representation in the unit
hypercube [0, 1] n (Balachandran et al., 2007). h ese detectors are evolved applying
a “structured genetic algorithm” (st. GA) with a niching technique for guiding the
search. A st. GA is a particular form of evolutionary algorithm, which incorporates
redundant genetic material controlled by a gene activation mechanism (Dasgupta
and McGregor, 1994). It utilizes multilayered genomic structures (hierarchical
chromosome) in which genes can be either active or passive (see Figure 4.15). An
activation mechanism enables and disables the encoded genes, and high-level genes
activate or deactivate sets of low-level genes. h e redundancy is used to maintain
genetic diversity to explore diff erent areas of the parameter space.
In this work, a structured GA gene with a two-level representation is used,
where the level 1 gene set holds the control information that either activates or

Control

Level 1

bits

Level 2

GS 2

…. GS n

GS 1

…

… …

g 11 g 12 … g 1k g 21 g 22 … g 2m g n1 g n2… g nr

Figure 4.15 Generalized representation of a chromosome with n different

gene sets.