The channel allocation and power control problem for multiple control loops over interference channels is investigated. A scheduler allocates the wireless channels and controls transmitting power based on observed channel conditions. The strategy can guarantee a QoS performance and minimize the average transmitting power simultaneously. The problem is formulated as an optimization problem and is solved by an ascent subgradient method. An offline algorithm to find the optimal allocation policy is put forward. Whereas, the algorithm is hard to be implemented in practice for requiring the channel distribution. Therefore, a stochastic subgradient ascent algorithm which is based on current observed channel state is proposed. It is proved that the stochastic optimal value converges almost surely to the optimal value. A numerical example illustrates the efficiency of the proposed method.