On static and dynamic code assignment policies in the OVSF code tree for CDMA networks

We introduce two new policies, static code assignment (SCA) and enhanced dynamic code assignment (EDCA), for assigning OVSF (orthogonal variable-spreading-factor) codes in heavily-loaded code-division multiple access systems that serve users with multiple data rates. The goal is to maximize the average throughput of the system. We present a numerical analysis based on a basic queueing model and compare the proposed techniques with the existing DCA scheme. The SCA assigns and fixes a partial set of OVSF codes in the binary code tree so that the average throughput is maximized for randomly arriving call requests. The configuration of the best partial set of OVSF codes depends on the ratios of the mixed call arrival rates. When the system becomes heavily loaded, SCA begins to outperform DCA in terms of average throughput. Another advantage of the static assignment is that it does not require any complex reassignments of codes to support a newly-arriving call. The EDCA resorts to a stationary dynamic call admission policy that accepts or rejects a newly arriving call on the basis of the state in the Markov decision process modeling call arrivals and departures. EDCA allows code reassignment to maximize the spectral efficiency. We derive an optimal EDCA policy through linear programming to solve for the optimal policy of the Markov decision process. Comparing DCA and EDCA, it is indicated that the greedy policy of DCA may not be optimal in terms of the long-term average rate of the total data served. Simulation and numerical evaluation confirms that EDCA performs better than DCA as the traffic load increases and high-rate calls become dominant.