ProtoG: a Computational Transgenetic Algorithm

Darwinian evolution has been the most usual metaphor to design evolutionary algorithms. Nevertheless, some researchers criticize the imperfection of the darwinian metaphor in solving several microbiological phenomena and in modelling evolution of species[1]. Genetic Algorithms, GA, introduced by Holland[6], are based on the darwinian paradigm and are successful in finding approximate solutions to many NPhard problems. However, there are several cases for which GAs fail. GAs use just the extra-cellular approach, that is, reproduction. On the other hand, DNA changes may also result from the intracellular flow and from epigenetic rules. In order to build on proteins, DNA uses a mechanism distinct to the one used on the creation of new individuals. Protein formation depends on the interactions that occur into the intracellular context, involving several agents. Furthermore, the information imprinting process is double way [4]. On the same way DNA prints information into messenger particles, it can also be influenced by such agents. Interactions in the intracellular context can affect reproductive cells and be inherited by offspring. One of the main characteristics of the intracellular flow is that it represents an intense and significative way to carry on physical-chemical interferences on the DNA. Several adaptation and evolutive tasks take place exclusively in the intracellular context [13]. Furthermore, recent researches show that genes and culture are inherently linked. Individuals evolve both by anatomical and behavioral selection. Rules that cause anatomical and behavioral elements come together are called epigenetic. Memes are the elements of cultural concepts. The term meme arose in the 1972 Richard Dawkins’ book The Selfish Gene to refers to unities of cultural transmission in analogy to the role of genes in biological evolution [3]. A meme is an idea like cooking meat before eat it. Memes may be transferred horizontally or from younger to older people. Memes can operate in large scales, helping to take care of the survival of larger groups of memes. Considering memes and epigenetic rules in evolutionary computation is a challenge [2]. Initially, in the computational context, memes were associated to exogenous information information out of the extra-cellular process such as, local search in GAs [11]. One of the main contributions of the transgenetic paradigm is that it is able to explore both environmental and cultural dimensions of the evolutionary process. A meme, in this work, is any proposal to construct a set or block of genes (building blocks). Therefore, Computational Transgenetic, CT, algorithms are designed to consider also the intracellular and epigenetic contexts on the evolutionary process. The next section presents CT and two of its transgenetic agents. Computational results of applying ProtoG to the Quadratic Assignment Problem are reported on Section 3. Finally, Section 4 presents some conclusions.