Genetic Programming as an approach to the problem of synthesis

Many problems do not have a direct solution in the form of a known algorithm or program to solve such a problem. These problems include, for example, the designing of electrical circuits and producing robots capable of locomotion. These are all part of a greater problem: the problem of synthesis. How can you make a computer design circuits and produce robots for you? In general, how can you automate the synthesis of anything that normally takes a great deal of human ingenuity? How a program should look to e.g. make robots walk is not known in advance, however the desired result is known. Genetic Programming provides a method to produce such algorithms or programs using evolutionary techniques applied iteratively to a population of programs to gain newer, more fitting programs which show the correct output. This paper gives an overview on the history of and developments in Genetic Programming. It describes Genetic Programming itself and where it fits among other Evolutionary Computation methods. Genetic Programming was studied and used by J. Koza which used GP for designing electrical circuits, GP is also used by H. Lipson to design the controller and morphology of robots