基于粒子群和鸟群优化的群智能算法研究.doc

摘要无论在科学计算还是工程应用中, 最优化问题都是非常重要的研究课题. 作为高效并行的启发式搜索算法, 粒子群和鸟群优化算法在求解最优化问题时都展现了其自身的优越性, 但仍然存在缺陷. 本文针对粒子群和鸟群优化算法收敛速度慢、寻优精度低等问题进行研究, 以此来弥补两种算法的不足之处. 本文的主要研究内容如下：1. 对粒子群和鸟群优化算法进行概述, 分析两种算法的优缺点, 并对粒子群和鸟群优化算法各提出了两种改进策略.2. 针对粒子群优化算法存在过早收敛、容易陷入局部最优等问题, 提出了两种改进策略. 对粒子群优化算法中的参数进行自适应调整, 平衡了算法的局部和全局搜索能力；引入混沌动态权重, 对粒子群优化算法中的更新公式进行改进, 加快了算法的收敛性能；建立机器人路径规划模型, 将改进的粒子群优化算法用于求解机器人路径规划问题, 通过数值试验验证了算法的有效性.3. 针对鸟群优化算法在求解高维复杂优化问题时存在收敛速度慢、寻优精度低等问题, 提出了两种改进策略. 将惯性权重引入鸟群优化算法中, 修正鸟群觅食策略, 增强了算法的多样性；引入云理论及均值概念, 对鸟群优化算法的更新公式进行改进, 加快了算法的搜索性能；建立农产品冷链物流配送路径优化模型, 将改进的鸟群优化算法用于求解农产品冷链物流配送路径优化问题, 通过数值试验验证了算法的可行性.4. 在多目标粒子群优化算法中, 为了弥补多目标粒子群算法的局部搜索能力和均匀性的不足, 引入一种竞争机制策略快速搜索非支配解, 对粒子的参数进行动态调整, 并且在算法后期引入时变高斯变异, 增强了算法的均匀性. 数值试验表明, 算法在收敛性和多样性方面都有所提升.5. 将鸟群算法应用于求解多目标优化问题, 提出了一种多目标鸟群优化算法. 引入克隆免疫策略, 增强了算法的随机性和多样性. 数值试验表明, 多目标鸟群优化算法具有较好的收敛性和多样性.关键词：粒子群优化算法, 鸟群优化算法, 多目标优化, 局部最优1AbstractIn scientific computing and engineering applications, the optimization problem is a very important research subject. As a highly efficient parallel heuristic search algorithm, particle swarm and bird swarm optimization algorithms in solving optimization problems are showed the superiority of its own, but there are still defects. In this paper, We studied the problems of slow convergence speed and low optimization accuracy of particle swarm and bird swarm optimization algorithms in order to make up for the shortcomings of the two algorithms. The main research contents of this paper are as follows:1. We summarized the particle swarm and bird swarm optimization algorithms, and alao analyzed the advantages and disadvantages of the two algorithms. Meanwhile we proposed two improved strategies for the particle swarm and bird swarm optimization algorithms respectively.2. In view of the problem that the particle swarm optimization algorithm has premature convergence and it is easy to fall into local optimum, we proposed two strategies to improve. The parameters in the particle swarm optimization algorithm are adjusted adaptively to balance the local search and global search capability of the algorithm. By introducing chaotic dynamic weights, we improved the updating formula of particle swarm optimization algorithm and accelerated the convergence performance of the algorithm. By establishing a robot path planning model, so that we improved particle swarm optimization algorithm to apply to solve the robot path planning problem. And the effectiveness of the algorithm is verified by numerical experiments.3. Aiming at the problems that the bird swarm optimization algorithm has slow convergence rate and low optimization accuracy in solving high-dimensional complex optimization problems, we also proposed two improved strategies. The inertia weight is introduced into the bird swarm optimization algorithm to modify the foraging strategy and enhance the diversity of the algorithm. By introducing the cloud theory and the concept of mean value, we improved for updating formula of the optimization algorithm, and accelerated the search performance of its algorithm. By means of establishing an optimization model of agricultural products cold chain logistics distribution path, we will improve bird swarm optimization algorithm is used to solve the optimization problem of cold chain logistics distribution path of agricultural products. Finally, the feasibility of the algorithm is verified by numerical test .4. In order to make up for the shortage of local search ability and uniformity of the multi-objective particle swarm optimization algorithm, we introduced a competition mechanism strategy is used to quickly search for non-dominant solutions and dynamically adjust the parameters of particles in the multi-objective particle swarm optimization algorithm. And introduced time-varying gaussian variation in the later stage of the algorithm. Numerical experiments show that the algorithm in convergence and diversity are improved.5. In this paper, in order to applied the bird swarm algorithm to solve the multi-objective optimization problem, we proposed a multi-objective bird swarm optimization algorithm, and introduced the clone immunity strategy to enhance the randomness and diversity of the algorithm. Numerical experiments show that the multi-objective bird swarm optimization algorithm has good convergence and diversity.Keywords: Particle swarm optimization algorithm, bird swarm optimization algorithm, multi-objective optimization, local optimizationIII目录摘要 IAbstract II第一章 绪论 11.1 课题研究的背景和意义 11.2 课题的国内外研究现状 21.2.1 粒子群优化算法 21.2.2鸟群优化算法 31.2.3 多目标粒子群优化算法 31.2.4 多目标鸟群优化算法 31.3 课题研究的目的 41.4 本文研究的主要内容和章节结构 41.4.1 本文研究的主要内容 41.4.2 本文的章节结构 5第二章 粒子群和鸟群优化算法概述 62.1 引言 62.2 粒子群优化算法 62.2.1 算法基本思想 62.2.2 算法原理 62.2.3 算法参数分析 62.2.4 算法基本流程 72.3 鸟群优化算法 82.3.1 算法基本思想 82.3.2 算法原理 82.3.3 算法参数分析 92.3.4 算法基本流程 102.4 本章小结 10第三章 改进的粒子群优化算法及其应用 113.1 引言 113.2 基于自适应策略的粒子群优化算法 113.2.1 自适应策略 113。