Optimal Joint Handoff and Code Assignment in CDMA Cellular Networks

Abstract

In this paper, we formulate the joint the handoff and code assignment problem in a CDMA cellular network as a reward/cost optimization problem. The probabilistic properties of the signals and of the traffic in the cells involved in the handoff are used to define a Markov decision process. The optimal policy for a hard handoff or soft handoff decision and channel (CDMA code) assignment is obtained by minimizing a cost function consisting of the weighted sum of the switching cost and the blocking rates of traffic in the two cells subject to a hard constraint on the bit error rate or the outage probability of the signal received by the mobile. A value-iteration algorithm is applied to derive the optimal policy. The performances of the optimal policies for joint hard or soft handoff and code assignment are evaluated and compared to that of the direct admission/threshold handoff policies. It is established that The optimal soft handoff policy results in the smallest expected number of handoffs compared to the other policies but in larger average blocking rate than the hard handoff policy. Both the optimal hard and soft handoff policies outperform by a significant margin the direct-admission/fixed-threshold policies.