Conclusion 

Producing the GGAT has been a good learning experience. As the research in genetic algorithms was done, the author became so fascinated with them that further study in this field was considers.

During the course of implementation certain issues came to light that would have enabled a radical re-design of the system. If the project had to be done again, less time would have been wasted on programming language choice, since that took a reasonable time of this project. LISP was initially considered as implementation language, but the learning curve proved to be too steep for the time provided.

Because of lack in time some features were not implemented that would have made the GGAT a better tool. The lack of error recovery during the execution of genetic algorithms, for example, should have been considered at earlier stages so that enough time was provided for its implementation. However these features can be easily added to the system, as well as some other features mentioned previously later on.

The main drawback of the system is the limited number of example problems implemented with it. Genetic algorithms are used it many area and mathematical function optimisation is just one of them.

Overall the aim of the project was accomplished and a general enough, genetic algorithm tool that can be reused for a variety of genetic algorithm implementations was developed. A reasonably easy to use GUI provided the front end of the system and an API provided for reuse of the genetic algorithm engine.

However all the work done for this project hardly scratch the surface of the genetic algorithms field, and much further work can be done on this project.