mmunication in the other direction. Typically, servers can and must send much more data to clients than clients send back to the servers. Examples include wireless networks with mobile clients, cable and satellite systems, push based technology used for broadcasting "heavily" used web pages etc. A common theme in all these applications is the use of broadcast as a means of information dispersal: the server has n pages which the clients need; the server can broadcast a fixed number of pages in each time slot; an arbitrary client wishing to access a particular page listens to the broadcast channel until its page appears on the channel. Given the access pattern (hit probabilities) of the n pages, an efficient broadcast schedule is one which minimizes the expected waiting time of the clients. We formulated a periodic scheduling problem [2] which models this broadcast application. Surprisingly t

uncertainty