Superframe Construction for Wireless Networks With Stochastic Demands

The link schedule of a wireless network must ensure links receive sufficient transmission opportunities or time slots to meet their offered load. To date, existing link schedulers assume fixed link load. In practice, link load is likely to vary, meaning the computed schedule or superframe will have unnecessary idle times. In this letter, we propose to use stochastic programming (SP) to generate a superframe comprising of a Time Division Multiple Access (TDMA) and a random access part. Advantageously, it sizes both parts according to traffic distribution. We show how it can be used to derive a superframe for multi transmit (Tx) or receive (Rx) wireless networks. We show via numerical results the efficacy of our approach in reducing idle times and collisions given random demands.