In this article, the spectrum trading problem between primary users and secondary networks is
investigated. The secondary network requests multiple channels with the targeted availability to satisfy its
users' demands. Due to the uncertainty about the channels availability, stochastic optimization techniques
are adopted to nd the optimal set of channels for each secondary network for the lowest cost. Two different
constraints on the secondary demand are dened. The rst one is when the throughput has to be fully satised
for a certain percentage of time, and the second one is when the expected value of the throughput has
to exceed a certain percentage of the requested one. Also, the possibility for channel subleasing among
the secondary networks is investigated to reduce the demand shortage. The results show that demanding
simultaneous channels increases the cost as it reaches up to 20% higher than if the same resources were
requested individually. Also, channels subleasing reduces the demand shortage probability and increases
the achieved throughput, especially at low value of requested demand. In this case, the demand satisfaction
probability increases by around 30% while the achieved throughput increases up to 40% compared to the
scenario where channels subleasing is not allowed.