Behavioral programming: a broader and more detailed take on semantic GP

In evolutionary computation, the fitness of a candidate solu-tion conveys sparse feedback. Yet in many cases, candidate solutions can potentially yield more information. In genetic programming (GP), one can easily examine program behav-ior on particular fitness cases or at intermediate execution states. However, how to exploit it to effectively guide the search remains unclear. In this study we apply machine learning algorithms to features describing the intermediate behavior of the executed program. We then drive the stan-dard evolutionary search with additional objectives reflect-ing this intermediate behavior. The machine learning func-tions independent of task-specific knowledge and discovers potentially useful components of solutions (subprograms), which we preserve in an archive and use as building blocks when composing new candidate solutions. In an experimen-tal assessment on a suite of benchmarks, the proposed ap-proach proves more capable of finding optimal and/or well-performing solutions than control methods