【作者】 彭倩；

【导师】 何蓉；

【作者基本信息】 西南交通大学 ， 通信与信息系统， 2012， 硕士

【摘要】 下一代无线通信系统需要支持高速率传输的无线多媒体业务,由于无线资源的稀缺性,如何合理有效利用无线资源是移动通信领域研究的重点。目前无线资源管理方面的研究主要包括资源控制、资源分配和资源调度三个方面。本文的研究重点是资源分配,通过改善OFDMA系统中的资源分配算法提高系统的性能。在下一代无线通信中,由于业务量的增大,能源的消耗也随之增加,因此需要提高系统效率降低整体能耗,实现绿色运营。本文从能效的角度考虑,通过自适应资源分配尽量最大化系统能效。目前OFDMA系统自适应资源分配算法中,最优算法的复杂度很高。本文以最小化系统总功率为优化目标,提出了一种复杂度较低的子载波-比特同时分配次优解算法。本算法不限制每个用户的子载波数目和子载波集合,在分配子载波的同时分配比特。本算法首先将用户的最小传输速率需求作为用户的需求因子,需求因子越大则表示用户对子载波和比特的需求程度越大,按照用户需求因子的大小依次为每个用户分配1条子载波和1bit；然后再将用户最小传输速率需求与当前时刻传输速率之比作为用户的需求因子,该比值越大则需求程度越大,根据需求程度的高低依次为用户分配子载波和比特,直到所有用户的最小传输速率需求被满足或者所有子载波和比特被分配完毕为止。在分配过程中,用户的需求因子是动态变化的。通过MATLAB仿真将本文算法与BABS-RCG算法和BABS-ACG算法进行了对比,结果表明本文算法的系统总功率和算法复杂度均得到了一定改善。此外,本文还研究了OFDMA系统中以能效为优化目标的自适应资源分配问题,采用了先子载波分配后比特分配的思想来实现最大化系统能效。由于以系统能效为优化目标的自适应比特-功率分配问题是一个非线性问题,不能通过用户之间的依次叠加来计算。因此,本文利用粒子群算法来解决自适应比特-功率分配问题。本算法中将多用户的比特分配方案作为粒子,通过计算系统能效来评价每个粒子的适应度,经过不断迭代寻找到最优的比特分配方案。通过MATLAB仿真,将本文算法与遗传算法进行了对比,结果表明粒子群算法的寻优能力和收敛性都要优于遗传算法,而且在用户的最小传输速率需求刚被满足时,系统的能效最大。更多还原

【Abstract】 The next-generation of wireless communication systems need to support high data rate wireless multimedia services. Due to the scarcity of radio resource, It is the key research point in the field of mobile communications that how to use the radio resource efficiently. Currently, researches of radio resource management have mainly focus on three aspects, namely, resource control, resource allocation and resource scheduling. This paper aims at the problem of resource allocation and intending to increase the system performance through improving the algorithm of resource allocation in the orthogonal frequency division multiple access (OFDMA) systems. In the next-generation wireless communication systems, the consumption of energy is increasing with the increment of traffic, so it is important to improve the efficiency of the systems and decrease the overall energy consumption for green operation. This paper attempts to maximize the energy efficiency via executing adaptive resource allocation algorithm.The current research works in adaptive resource allocation of OFDMA systems usually adopt optimal algorithm with high complexity. This paper proposes a low complexity sub-optimal algorithm which allocates the sub-carrier and bit simultaneously, and formulates the optimal goal as to minimize the overall transmit power. The proposed algorithm has no limit for the number of sub-carriers and the sub-carrier set of every user. Firstly, we introduce the minimum data rate requirement as the user demand factor. The bigger the user demand factor is, the larger the requirement degree of user for sub-carrier and bit is. We allocate one sub-carrier and one bit to each user according to the user demand factor. Then, we use the ratio of the minimum data rate requirement and the instant rate as the user demand factor. Similarly, the bigger the user demand factor is, the larger the requirement degree of user for sub-carrier and bit is. We allocate the sub-carriers and bits to users according the sequence of user demand factor, until the minimum data rate requirements of all users are satisfied or all sub-carrier and bits are exhausted. Note that the user demand factor is dynamic during the process of allocation. We compare the performance of the proposed algorithm with BABS-RCG algorithm and BABS-ACG algorithm by matlab simulation. Simulation results indicate that the system power and the computational complexity of the proposed algorithm are significantly improved.In addition, this paper addresses the adaptive resource allocation problem based on optimization of energy efficiency. We introduce the idea of sub-carrier allocation firstly and bit allocation secondly to maximize the energy efficiency. Since the problem of adaptive bit-power allocation with the goal of energy efficiency optimization is a nonlinear problem, it is impossible to solve the problem by ordinal iterative addition among users. This paper uses the Particle Swarm Optimization (PSO) to solve the adaptive bit-power allocation problem. We adopt the bit allocation scheme of multi-user as particle, and evaluate the fitness degree of every particle by computing the energy efficiency of system. In order to obtain optimal bit allocation scheme, we execute continuous iteration. We compare the performance of the proposed algorithm with Genetic Algorithm (GA) by matlab simulation. Simulation results show that both the seeking optimization ability and the convergence performance of PSO are better than GA. Furthermore, it will reach the optimization point of energy efficiency only when the data rate requirements of each user are just meet.更多还原