Power and Spectrum Allocation in D2D Networks Based on Coloring and Chaos Genetic Algorithm

To tackle the current issue of spectrum resource scarcity, Device-to-Device (D2D) communication is considered as an important technology for sharing spectrum with cellular users, which enables high-speed and convenient services. In this paper, we consider the spectrum and power allocation problem for D2D communication in underlaying cellular network. In particular, we focus on a D2D network, where spectrum resource is shared among all D2D users. Unlike existing works that mainly focus onD2D users sharing spectrum with cellular users, we analyze not only the spectrum resources allocation but also the optimization of the transmission power for each D2D transceiver pair to maximize D2D system capacity maximum. We first formulate the network model and signal-to-interference-plus-noise ratio (SINR) by some stochastic geometry knowledges. In addition, we combine the spectrum resource distribution with power selection and transform this problem into a mixed integer non-linear program(MINLP) to maximize system capacity. Inspired by existing intelligent algorithms, we propose a heuristic chaos genetic algorithm associated with four color theorem to solve this problem. In order to prove the feasibility and efficiency of this algorithm, we compare this problem with brute-force search algorithm in simulation results. It shows that the proposed chaos genetic algorithm combine with graph coloring method can approach system capacity optimal but algorithm complexity is greatly reduced. We also verify the system parameter influence of the system capacity in this network.