The field of biological evolution brought a new age in adaptive computation (AC). Among different evolutionary computation approaches, Genetic Algorithms (GA) are receiving much attention both in academic and industries. Genetic algorithms are general-purpose search procedures based on the mechanisms of natural selection and population genetics. Genetic algorithm-based tools have started growing impact in companies - predicting financial market, in factories - job scheduling etc. with their power of search, optimization, adaptation and learning. For the users of diversified fields, genetic algorithms are appealing because of their simplicity, easy to interface and ease to extensibility.
Despite their generally robust character, as the application increases, there found many domains where formal GAs perform poorly. Several modifications have been suggested to alleviate the difficulties both in the manipulation of encoded information and the ways of representing problem spaces. A number of different models, namely, Messy GAs and Genetic Programmings developed recently which addressed the representation issue of GAs.
Dipankar Dasgupta has been involved in the investigation of a more biologically motivated genetic search model - called the Structured Genetic Algorithm (sGA). The model uses some complex mechanisms of biological systems for developing a more efficient genetic search technique. Specifically, this model incorporates redundant genetic material and a gene activation mechanism which utilizes multi-layered genomic structures for the chromosome. The additional genetic material has many advantages in search and optimization. It mainly serves two purposes: primarily, it can maintain genetic diversity at all time during the search process, where the expected variability largely depends on the amount of redundancy incorporated in the encoding.
The following paragraphs summarize some aspects and advantages of Structured Genetic Algorithms:
One school of thought (Darwinian) believes that evolutionary changes are gradual; another (Punctuated Equilibria) postulates that evolutionary changes go in sudden bursts, punctuating long periods of stasis when very small evolutionary changes take place in a given lineage. The new model provides a good framework for carrying out studies that could bridge these two theories.
Structured GA's results to date are very encouraging, though there remain many issues for further investigation. It appears to an enhancement of the formal genetic model with a number of practical advantages. This approach has also received favorable attention in the field of evolutionary computation. However, the studies on structured GAs done so far are only the first step toward the broader goal of developing a more efficient genetic search. Further research to understand the behavior of the model and to determine its search properties is in progress.