Adaptive Load Balancing MDP-Based Approach of Two-Dimensional Spreading for VSF-OFCDM in 4G Next-Generation Cellular Communications

Abstract

The variable spreading factor-orthogonal frequency and code division multiplexing (VSF-OFCDM) system has been proposed as the forward link interface for achieving a high data rate in fourth-generation (4G) mobile communications. VSF-OFCDM allocates the orthogonal channelization codes of an orthogonal variable spreading factor (OVSF) code tree in 2-D spreading in the time and frequency domains but suffers from two disadvantages, namely, moderate utilization and low transmission quality. Moderate utilization is caused by code blocking, and the low transmission quality is due to the multicode interference from high channel loading in the time domain of the 2-D spreading. A tradeoff thus exists between code blocking and multicode interference. A lower code blocking means a higher interference. Achieving high utilization while also providing high transmission quality is a critical issue that should be addressed in 4G VSF-OFCDM. Therefore, this paper proposes a 2-D spreading approach that is based on an adaptive load balancing with Markov decision process (which is denoted by ALM). The ALM approach consists of three phases. First, for balancing the channel load, the adaptive 2-D spreading phase is proposed to select the 2-D spreading combinations, which have low channel load and high frequency diversity, as the candidates. Second, a dynamic recombination of the 2-D spreading is presented to decrease the channel load while supporting the transmission quality from high frequency diversity if the channel load of a time-domain code exceeds the defined channel load threshold. Third, the cost-based Markov decision process (MDP) code-selection approach is adopted to minimize the code blocking. The MDP models the state of the OVSF code tree as an MDP and then defines the costs for all the candidate channelization codes. The MDP phase selects the least-cost channelization code as the optimal solution if several 2-D spreading combinations can satisfy the channel load limitation. Additionally, the time complexity of the ALM is analyzed. The numerical results indicate that the proposed approach outperforms the other approaches in transmission quality ratio, fractional reward loss, and total number of reassignments.