An ant–colony approach for the design of optimal chunk scheduling policies in live peer–to–peer networks