Incremental Evaluation in Genetic Programming

Often GP evolves side effect free trees. These pure functional expressions can be evaluated in any order. In particular they interpreted from the genetic modification point outwards. Incremental evaluation exploits fact that: highly evolved children semantic difference between child and parent falls with distance syntactic disruption (e.g. crossover point) reach zero before whole has been interpreted. If so, its fitness is identical to (mum). Considerable savings bloated binary tree runs are given by exploiting population convergence existing GPquick data structures, leading near linear O(gens) runtime. With multi-threading SIMD AVX parallel computing a 16 core desktop deliver equivalent of 571 billion operations per second, giga GPop/s. viewed via information theory as evolving smooth landscape software plasticity. Which gives rise resilience source code changes. On average mixture 100 +, −, $$\times $$ (protected) $$\div nodes remove test case effectiveness at exposing changes so fail propagate infected errors.