DGP: How To Improve Genetic Programming with Duals

Abstract

In this paper, we present a new approach, improving the performances of a genetic algorithm (GA). Such algorithms are iterative search procedures based on natural genetics. We use an original genetic algorithm that manipulates pairs of twins in its population: DGA, dual-based genetic algorithm. We show that this approach is relevant for genetic programming (GP), which manipulates populations of trees. In particular, we show that duals can transform a deceptive problem into a convergent one. We also prove that using pairs of dual functions in the primitive function set, is more efficient in the problem of learning boolean functions. Here, in order to prove the theoretical interest of our approach (DGP: dual-based genetic programming), we perform a numerical simulation.