【作者】 杨传鑫；

【导师】 颜学峰；

【作者基本信息】 华东理工大学 ， 控制科学与工程， 2011， 硕士

【摘要】 在智能优化算法寻优过程中,算法参数是否合适、是否能随着进化搜索进展自适应调整,将直接影响算法的最终寻优性能。模糊逻辑是近年来提出的一种能融合专家先验知识的自适应调整策略,本文将模糊控制引入两类智能优化算法(即遗传算法与粒子群算法),通过融合算法参数调整的先验规律,实现算法参数的自适应调整,仿真研究与实际应用表明算法寻优性能获得提高。针对遗传算法控制参数的自适应调整,提出一种改进模糊自适应遗传算法(FAGA)。FAGA算法的交叉概率和变异概率采用模糊推理动态调整；同时,采用免疫原理改进遗传算法的选择操作,增强种群的多样性。通过标准测试函数仿真测试,FAGA寻优性能优于两个经典模糊遗传算法。最后,将FAGA应用于氧化动力学模型参数估计。结果表明：FAGA寻优结果明显优于Powell,以及其他两种改进的遗传算法。针对粒子群算法中惯性权重自适应调整,提出一种改进模糊粒子群算法。改进算法采用模糊控制器自适应控制惯性权重,并且在算法陷入局部最优的时候加入随机速度,提高全局寻优概率。标准测试函数的仿真测试表明：改进算法可以有效克服早熟收敛,并具有较好的收敛速度；寻优性能优于标准PSO算法。同时,提出一种自适应惩罚函数解决约束性优化问题,提出的惩罚函数能较好的处理等式约束与不等式约束都多的问题。仿真实验结果表明：提出的惩罚函数性能优秀。最后,通过改进模糊粒子群算法、以及提出的惩罚函数优化水系统模型,获得良好结果。更多还原

【Abstract】 Some factors have directly effect on the performance of intelligent optimization algorithms. For example, whether the algorithm parameters are appropriate, or whether the algorithm parameters are adjusted to obtain adaptive value during the optimization process. Fuzzy logic, which is a kind of self-adaptation adjust tactics with a prior knowledge of experts, is employed for adjusting the algorithm parameters. This article will propose two types of adaptive intelligent optimization algorithm based on fuzzy control of the algorithm parameters (i.e. genetic algorithms and particle swarm optimization). In order to adjust algorithm parameters, the rules, which integrate a priori knowledge, are used. Results of the simulation show that optimization performance is improved.A novel fuzzy-based adaptive genetic algorithm (FAGA), which has adaptive control strategies for parameters of genetic algorithms, is proposed. In FAGA, crossover probability and mutation probability are adjusted dynamically by fuzzy inferences, which are based on the heuristic fuzzy relationship between algorithm performances and control parameters. And, immune theory is used to improve the selection operation of GA and increase diversity. The experiments show that FAGA can efficiently overcome shortcomings of GA, i.e. premature and slow, and obtain the better results than two typical fuzzy GAs. Finally, the result is satisfactory when the algorithm is used for the parameters estimation of reaction dynamics model. The optimal value obtained by FAGA is better than Powell and other two improved genetic algorithm.A new fuzzy-based adaptive Particle Swarm Optimization algorithm (FPSO), which uses Fuzzy Logic Controller to control inertia weight, and adds random velocity to increase the probability of global optimization when the algorithm falls into the local convergence, is proposed. Experimental results show that FPSO effectively overcomes slow and premature in unconstrained optimization problem, and has better ability to find the global optimum than the standard PSO. This paper also proposes a self adaptive penalty function for solving constrained optimization problems by applying FPSO. And, the proposed penalty function has the great ability for the optimized problem, which not just has more equality constraints but also more equality constraints. Experimental results demonstrate that FPSO integrated with the proposed penalty function has better performance than other three kinds of penalty function for solving the constrained problem. Finally, FPSO integrated with the proposed penalty function was used to optimize the water system, and the satisfactory result was obtained.更多还原