Skip to Main Content
A novel joint power and rate adaptation scheme for DS/CDMA networks is proposed using information from the TCP state machine. The impetus behind this cross-layer optimisation approach is to satisfy the varying transmission rates of TCP flows which depend on the congestion window and round-trip time. Initially, a new family of objective functions is defined that encompasses TCP based information. Then by solving the proposed bi-objective optimisation problem using the ε-constraint method the Pareto front is calculated, which gives insight into the trade-off between the desired transmission rate and the corresponding transmitted power. Using a polynomial description of the Pareto front, the solution with the minimum L2-norm from a defined utopia point is selected as the optimum trade-off between power consumption and required rate transmission, even though different heuristics can be used. Although the Pareto frontier has been calculated using derivative based optimisation techniques, prominent paradigms of knowledge-based techniques can increase the computational efficiency. Previously proposed schemes allocate resources based on lower-layer information, such as channel conditions and link-layer buffer occupancy. The novelty and prerogative of this cross-layer optimisation approach is that critical transport layer information is integrated into the design of a power and rate adaptation algorithm for DS/CDMA networks.
Date of Publication: 24 Dec. 2004