| Summary: | 碩士 === 國立成功大學 === 數學系應用數學碩博士班 === 98 === In this thesis, we focus on the optimal congestion control of the wireless ad-hoc network. First, we formulate the multi-hop cross layer resource management problem by finding the jointly optimal power allocation and end-to-end communication rates of an ad-hoc network. Secondly, by the analysis of the congestion control model, non-convex capacity formulas and the minimax fractional objective are the two major technical difficulties for solving the model. However, the fairness of resource allocation is the key idea for solving the minimax fractional problem.
Our idea is to design a locally linear approach method (LLAM) to approximate the non-convex capacity and apply the modified Dinkelbach entropic regularization (MDER) method to cope with the minimax fractional programming. In addition, by adding some additional constraints, we can control the fairness of resource allocations to achieve steadiness in the entire system.
Finally, our numerical results indicate that the congestion can be relieved efficiently.
|
|---|
