Algorithm Research On OFDM Resource Allocation

LTE system’s core transmission technology is OFDM (orthogonal frequency division multiplexing) technology. In OFDM system, Broadband carrier is divided into a number of frequency orthogonal sub-carriers for data transmission by Fast Fourier Transform (FFT), each user can be assigned to a plurality of subcarriers. All subcarrier’s fading conditions are independent of each other for each user channel, therefore how to allocate the subcarrier and how to determine the power of each subcarrier are very important for improving the transfer rate, system capacity and wireless communication resource utilization.In the OFDM resource allocation mathematical model, because the model is too complex, and the model is a hybrid optimization algorithm, it is very hard to optimization of the model. Most algorithms are based on the framework of the two-step method. The first step is to allocate subcarrier according to certain criteria, and the second step is to allocate the power, and optimize the power allocation model. The two-step method can significantly reduce the complexity of the algorithm, and improve the algorithm for rate. However, due to the sub-carrier assignment and power allocation are not independent in OFDM resource allocation model, two-step method’s performance is not very good. However, if the subcarrier allocation and power allocation optimization are optimized at the same time, the algorithm complexity would be too high, it is not meet the real time requirements of mobile communication system. If in the iterative process, subcarrier allocation and power allocation turns to be optimized, and the information of subcarrier and power allocation interact with each other, it is expected to improve the algorithm accuracy without increasing the complexity of the algorithm.In this study, an evolutionary algorithm for OFDM resource allocation based on simulated annealing algorithm is proposed. This algorithm generates a subcarrier allocation program firstly, and the evolutionary algorithm is used to solve the nonlinear equation of resource allocation. Whenever the algorithm runs G (G is a constant) generations, the sub-carrier allocation scheme will be changed slightly. The new program will be compared with the old one. If the former is better than the latter one, the new program will replace the old one. Otherwise, the Boltzmann acceptance probability will be used to decide which program is selected. Repeat the above process until the termination condition is reached. Simulation results demonstrate that the proposed resource allocation algorithm can achieve perfect performance, mean-while efficiently improving the system rate.An evolutionary algorithm for OFDM resource allocation based on KKT conditions is proposed in the3th section. The algorithm makes full use of the feature that OFDM resource allocation problem is a mixed optimization problem, using evolutionary algorithm to optimize discrete variables and using KKT conditions to optimize continuous variables. Firstly, this algorithm generates a set of subcarrier allocation programs as a population of the evolutionary, and each subcarrier allocation program is an individual. After the subcarrier allocation program is determined, the original OFDM resource allocation model becomes a nonlinear optimization problem with equality and inequality constraints. KKT conditions are used to optimize the nonlinear optimization problem, and a stagnation point can be obtained. The transfer rate of the system can be gotten through the power allocation result obtained by KKT conditions and the subcarrier allocation program. The transfer rate of the system is regarded as the fitness of the individual. Algorithm proposed in this study can make the information of subcarrier and power allocation interact with each other. Computer experiments show the proposed algorithm is effective.