Size Fair Tree Crossovers

Size fair crossover genetic operator for tree based genetic programming is described and tested. It produces considerably reduced increases in program size and no detrimental effect on GP performance. GP search spaces are partitioned by the ridge in the number of program v. their size and depth. A ramped uniform random initialisation is described which straddles the ridge. With subtree crossover trees increase about one level per generation leading to sub-quadratic bloat in length