Genetic Programming in C++: Implementation Issues
Mike J. Keith, Martin C. Martin · The MIT Press eBooks · 2003
Introduction In this chapter we explore the lower level implementation issues surrounding what we call the Genome Interpreter. Provided is example code from 5 test programs which were used to evaluate performance. Section 13.8 summarizes the results of these tests and discusses the trade-offs involved with the various implementations. For the upcoming discussion, what we call an interpreter specifies the following lower level aspects of the design: . the raw node representation . how a tree of nodes is represented . the method for evaluating an individual node . the method for evaluating the tree as a whole . the methods for (or methods to assist) those genetic operators which are dependent on the node or tree representation. A key point is that the interpreter specifies the node implementation which is the particular part of the platform-coding in which the overhead will be magnified. Therefore, the interpreter is the most crucial component in the overall design with respect t