论文部分内容阅读
基于模拟退火算法(SAA)和鲍威尔算法(PA)提出了一种新的混合优化算法——模拟退火鲍威尔算法(SAA-PA).该算法将模拟退火算法的随机搜索策略纳入到鲍威尔优化算法中,能使优化解不陷入局部最优从而获得全局优化解.该优化方法可以有效地解决多目标的优化问题,特别适用于拥有多个局部最优值的分布式多泵浦拉曼放大器(DMRA)的优化问题.仿真结果显示,在80km传输光纤上只要4个后向泵浦就能实现开关增益10dB,带宽大于80nm,增益平坦度小于1.25dB的平坦增益.与单纯的模拟退火算法的优化结果相比,在相同数目的泵浦条件下所得优化结果的增益谱特性有了显著的提高.同时,该方法可以较方便地仿真出信号增益带宽与泵浦数目的内在关系.
A new hybrid optimization algorithm called simulated annealing Powell’s algorithm (SAA-PA) is proposed based on SAA and Powell’s Powersave (PA) algorithm. This algorithm incorporates the random search strategy of simulated annealing algorithm into Powell’s optimization algorithm , So that the optimal solution can not get into the local optimum to obtain the global optimal solution. This optimization method can effectively solve the multi-objective optimization problem and is especially suitable for distributed multi-pump Raman amplifiers (DMRA) with multiple local optimums ) Simulation.The simulation results show that the flat gain of 10dB switching gain, bandwidth of more than 80nm and gain flatness less than 1.25dB can be achieved with only four backward pumps on the 80km transmission fiber.Compared with the simple simulated annealing algorithm Compared with the experimental results, the gain spectrum characteristics of the optimized results obtained under the same pump conditions have been significantly improved.At the same time, this method can more easily simulate the intrinsic relationship between the signal gain bandwidth and the pump number.