Random generation of Boolean functions with high degree of correlation immunity

In recent years a cryptographic community is paying a lot of attention to the constructions of so called resilient functions for use mainly in stream cipher systems. Very little work however has been devoted to random generation of such functions. This paper tries to fill that gap and presents an algorithm that can generate at random highly nonlinear resilient functions. Generated functions are analyzed and compared to the results obtained from the best know constructions and some upper bounds on nonlinearity and resiliency. It is shown that randomly generated functions achieve in most cases results equal to the best known designs, while in other cases fall just behind such constructs. It is argued that the algorithm can perhaps be used to prove the existence of some resilient functions for which no mathematical prove has been given so far.