Abstract: In this paper, a parallel analogue/digital implementation of the k-winner-take-all (k-WTA) function is implemented. A hardware and software realization of this neurocomputing system based on competitive learning is carried out for the purpose of controlling and managing water levels in a water shortage areas based on water need and not on water demand.