ABSTRACT
Genetic programming (GP) (Koza 1992) was first proposed by John R. Koza to construct programs solving given problems. Although there appear many GP variants (Oltean et al. 2009, O’Neill & Ryan 2001, Byrne et al. 2014 in press, Montana 1995, Sabar et al. 2013) such as Gene Expression Programming (GEP), Grammatical Evolution (GE), Strongly Typed Genetic Programming (STGP), etc. since then, their way to construct expected programs from given sets of components are almost the same. They all solve optimization problems with random search technologies. This means few knowledge or constrain can be utilized to optimize the evolution process, therefore encountering diculty in narrowing down the big search space.
