A chemical genetic algorithm
(CGA)
in which several types of molecules (information units) react with
each other in a cell is proposed. Translation from codons (short
substrings of bits) in DNA to amino acids (real value units) is
specified by a particular set of translation molecules created by
the reaction between tRNA units and amino acid units. This
adaptively changes and optimizes the fundamental genotype-phenotype
mapping
during evolution. Through the
struggle between cells containing a DNA unit and small molecular
units, the codes in DNA and the translation table described by the
small molecular units coevolve, and a specific output function
(protein), which is used to evaluate a cell's fitness, is optimized.
To demonstrate the effectiveness of the CGA, the algorithm is
applied to a set of deceptive problems, and the results by using the
CGA are compared to those by using a simple GA. It is shown that
the CGA has far better performance for the tested functions than the
conventional simple GA.