The "Clones" of the Incursive Genetic Algorithms

This paper deals with a singular comportment of an incursive algorithm. This algorithm has been primary developed to optimise a production process. It is closed to classical genetic algorithms because it uses these major stages determination of n random solutions, evaluation of the solutions (in our case with the simulation of production process), selection of the best solution and reproduction according to the biologic rules of crossing over. Nevertheless the differences between this algorithm and a classical genetic algorithm are : the use of incursion in the selection's stage and the singularity of the reproduction stage. The algorithm enables indeed to optimise the system configuration by optimising the laws of reproduction. The "crossing over" becomes a specific case among huge other configurations. Despite we expect that the algorithm will be able to adapt his reproduction laws to the specificity of the system studied, the comportment of the optimisation, due to the use of incursion is unexpected and unfortunately unable to solve our production matter. Nevertheless this comportment is very interesting to study the consequences of the use of incursion with a genetic algorithm.