粒子群优化算法(PSO)来源于鸟类的群体行为，是一种基于种群的随机搜索算法。粒子群优化算法的性能在很大程度上取决于参数的选择。其中，惯性权重是该算法的参数之一，用于平衡粒子群优化算法的探测和开发能力。为了提升算法的性能，PSO应较好地兼顾全局搜索能力与局部搜索能力。PSO算法具有收敛速度快、易于实现、参数调整少等优点，其有效的搜索策略使之具有潜力在各种应用中解决不同优化问题的方法。PSO在电力系统优化、过程控制、动态优化、自适应控制和电磁优化等领域有着广泛的应用。虽然PSO在求解许多优化问题时表现出良好的性能，但与大多数随机搜索方法一样，它也存在早熟收敛问题，尤其是在多模态优化问题中。虽然取得了显著的进展和丰硕的成果，但如何较好地平衡粒子群优化算法的探测和开发能力以确定复杂优化问题的高质量解决方案，仍然具有一定的挑战性。本文将局部搜索与粒子群算法相结合，以平衡算法的探测与开发能力，并将改进后的方法应用于许多优化问题和实际问题。本文的主要工作如下： (1)为了解决这一问题，本文提出了一种基于梯度局部搜索策略的自适应惯性权重粒子群优化算法(SIW-APSO-LS)。该算法旨在平衡全局搜索算法的探测能力以及局部搜索策略的开发能力。首先，采用自适应权重粒子群优化算法(SIW-APSO)对问题进行求解，在进化过程中对粒子的速度与位置进行迭代更新。其次，基于梯度的局部搜索策略注重于对搜索区域进行开发，在SIW-APSO之后执行精确搜索。实验结果证明了算法的有效性，在基准测试函数集合上取得了优于传统PSO方法的性能。 (2)为了解决传统算法收敛速度较慢的问题，本文进一步提出了一种新颖的基于群体的混合优化算法， 该算法融合了动态的多群粒子群优化算法以及引力搜索算法(GSADMSPSO)。该算法将主种群切割为多个较小的子种群，并通过提出新的邻域策略使各个子种群更加稳定。通过PSO的全局优化能力，每一个粒子的速度与位置信息均得到更新。使用引力搜索算法并结合动态的多种群算法能够有效地提高算法的探测与开发能力。在实验部分使用基准测试函数验证所提出的算法性能。实验结果表明， GSADMSPSO能够取得比基于引力搜索或PSO的算法更好的结果。 (3)特征选择是一项能够提高分类器的预测性能与降低问题复杂度的重要技术。目前，大量的研究已经采用启发式优化方法进行特征选择，且取得了不错的效果。本文提出了一种基于SIW-APSO-LS与C4.5分类器的特征选择方法。在该算法中，我们结合了基于梯度的局部搜索策略能够有效帮助开发搜索空间以及自适应权重粒子群优化算法能够收敛到搜索空间的最优解这两点优势进行特征选择。实验结果表明，在UCI数据库中16个标准数据集上，所提出的算法取得了比传统特征选择算法更优秀的结果。 (4)本文采用GSADMSPSO训练前馈神经网络(feedforward neural networks， FNNs)，通过是否能够避免局部最优问题以及相比于传统进化算法收敛速度是否能够加快这两方面验证其性能。将所提出的算法的性能与引力搜索算法及其变体在不同基准测试函数上进行比较。实验结果表明，该算法在收敛速度和避免据最小值等方面由于其他的训练前馈神经网络的方法。 关键词(Keyvords)：自适应粒子群优化；基于梯度的局部搜索；引力搜索算法；特征选择；前馈神经网络

 Particle swarm optimization (PSO) is a stochastic population-based algorithm motivated by intelligent collective behavior of birds. The performance of the PSO algorithm highly depends on choosing appropriate parameters. Inertia weight is a parameter of this algorithm which was used to a balance between the exploration and exploitation characteristics of PSO. PSO should have strong yet balanced exploration and exploitation capabilities to enhance its performance. The advantages of PSO algorithm include fast convergence toward the global optimum, easy implementation and few parameters to adjust. Its effective searching strategy makes it a potential method for solving different optimization problems in a wide variety of applications. PSO has been widely used in many fields such as power system optimization, process control, dynamic optimization, adaptive control and electromagnetic optimization. Although PSO has shown good performance in solving many optimization problems, it suffers from the problem of premature convergence like most of the stochastic search techniques, particularly in multimodal optimization problems. Although noticeable progress and fruitful achievements have been attained, successfully balancing the exploration and exploitation capabilities of PSO to determine high-quality solutions. This thesis will combine local search with PSO to balance exploration and exploitation capabilities, and apply the improved proposed method to many optimization problems and real-world problems. The main works of this thesis are listed as follows: (1)An improved self-inertia weight adaptive particle swarm optimization with a gradient-based local search strategy (SIW-APSO-LS) is proposed. This new algorithm balances the exploration capabilities of the improved inertia weight adaptive particle swarm optimization and the exploitation with the gradient-based local search strategy. The self-inertia weight adaptive particle swarm optimization (SIW-APSO) is used to search the solution. The SIW-APSO is updated with an evolutionary process in such a way that each particle iteratively improves its velocities and positions. The gradient-based local search focuses on the exploitation ability because it performs an accurate search following SIW-APSO. Experimental results have verified that the proposed algorithm performed well compared with other PSO variants on a suite of benchmark optimization functions. (2)A new hybrid population-based algorithm is proposed with the combination of dynamic multi swam particle swarm optimization and gravitational search algorithm (GSADMSPSO). The proposed algorithm divides the main population of masses into smaller sub-swarms and also stabilizing them by presenting new neighborhood strategy. At this time, by adopting the global search ability of proposed algorithm, each agent (particle) improves the position and velocity. The main idea is to integrate the ability of GSA with the DMSPSO to enhance the performance of exploration and exploitation of a proposed algorithm. In order to evaluate the competences of the proposed algorithm benchmark functions are employed. The experimental results have been confirmed a better performance of GSADMSPSO as compared to the other gravitational and PSO variants in terms of fitness rate. (3)Feature selection is an important task to improve prediction accuracy of classifiers and to decrease the problem size. Numerous methodologies have been presented to achieve feature selection using metaheuristic algorithms. In this study, an improved self-hybrid individual inertia weight adaptive particle swarm optimization with local search and combined with C4.5 classifiers for feature selection algorithm is proposed. In this proposed algorithm, we combine the gradient base local search with its capacity of helping to explore the feature space and improved self-hybrid individual inertia weight adaptive particle swarm optimization with its ability to converge to the best global solution in the search space. Experimental results have verified that the SIW-APSO-LS performed well compared with other current state of art feature selection techniques on a suit of 16 standard data sets obtained from the UCI repository. (4)In this study, GSADMSPSO is employed as new training methods for FNNs in order to examine the efficiencies of the algorithm in decreasing the problems of trapping in local minima and the slow convergence rate of current evolutionary learning algorithms. The performance of the proposed algorithm is compared with GSA and its variants on different suites of well-known benchmark test functions. The experimental results show that proposed algorithm outperforms to other variants training FNNs in terms of converging speed and avoiding local minima. Keywords Adaptive particle swarm optimization; gradient based local search; gravitational search algorithm; feature selection; feedforward neural networks