摘要
针对非线性系统的建模问题,借鉴种群划分和递阶进化的思想,设计了一种基于双层优化的多模型建模方法。该方法将多模型建模问题转化为一个双层优化问题,在上层采用多个种群对多模型的区域进行优化划分,在下层采用多个个体微粒对各个局部模型的参数进行寻优,从而有效避免了多个参数同时优化带来的局部最优问题。最后,通过微粒群优化算法对其进行求解,并通过一个仿真算例验证了该建模方法的有效性。
With reference to the multi-species partition and the hierarchy evolutionary, a bilevel optimization based multiple-model modeling method was introduced to deal with the modeling problem of nonlinear systems. The multiple-model modeling problem was transformed to a bilevel optimization problem, with the optimal partition being realized by means of the multi-species at the upper levels, and the parameters of each local model in the species being optimized at the lower levels, which effectively avoided the local minimum problem caused by the simultaneous optimization of multiple parameters. Finally, the Particle Swarm Optimization （PSO） algorithm was employed to solve the problem, and a simulation example indicated the effectiveness of the proposed modeling method.
出处
《计算机应用》
CSCD
北大核心
2009年第5期1261-1263,共3页
journal of Computer Applications
基金
总装备部武器装备预研基金资助项目(9140A04050407JB3201)
关键词
双层优化
微粒群算法
混合逻辑模型
多模型
优化建模
bilevel optimization
Particle Swarm Optimization （PSO） algorithm
mixed logical model
multiple models
optimization modeling
