The field of Genetic Programming (GP) is concerned with harnessing the power of simulated evolution to search massive expression spaces, so as to discover a functional mapping between a non-trivial set of inputs to an arbitrary output. The problems that GP are applied to are often NP-Complete and intractable by traditional means. This is achieved by maintaining populations of potential solutions, referred to as individuals. In standard GP, individuals are expression trees. When applying GP to a problem, one typically executes multiple runs, this is due in part to stochastic nature of GP, as each run executes differently, and its success or otherwise depends on a combination of the initial population and the random choices made thereafter.
展开▼