基于双因子改进型粒子群算法的混合滤波系统的多目标优化设计

Multi-objective Optimization Design of Hybrid Filter Based on Improved Particle Swarm Optimization Algorithm with Two Factors

大容量混合滤波系统的应用前景十分广阔,其滤波器组的参数优化,尤其是无源滤波器的参数设计对综合性能的影响很大。本文提出基于双因子改进型粒子群优化算法的混合滤波系统中无源滤波器组的多目标优化设计方法。通过将加速因子和交叉因子引入优化算法,针对无源滤波器的多目标优化,高效搜索参数空间以获得最优解。与以往的优化设计相比,带双因子的改进型粒子群算法采用自适应的惯性权重,交叉因子增加了粒子的多样性,克服了算法寻优过程中易局部收敛等问题,提高其全局搜索能力。而收缩因子可以加快PSO的运算速度,明显提高了算法的寻优速度。测试结果表明利用该算法设计的混合型滤波系统具有较强的鲁棒性。

The high-capacity hybrid filter will be applied very widely in the large ship. The optimization of parameters has great impact on its overall performance. In this paper, the multi-objective optimization method based on two-factor improved particle swarm optimization（PSO） algorithm is introduced in the design of passive filters group of the hybrid filter equipment, with the aim to resolve the current problems such as poor capacity and tardy response in optimization. With the application of acceleration factor and hybrid genes to multi-objective optimization, higher efficient search within the parameter space can be realized to obtain the optimal solution. Compared with the traditional optimal design, the improved PSO with two factors adopts self-adaptive inertia weight, which allows hybrid genes to diversify the particles and overcomes local convergence in the algorithm optimization process, and the capacity of overall search is enhanced consequently. Constriction factor can accelerate the computation speed of PSO,which improves the algorithm＇s searching speed greatly. Experiments indicate that filtering effects are improved and that the hybrid filters have good performance and robustness.