风力驱动优化算法及其应用研究

发布时间：2018-01-01 22:08

本文关键词：风力驱动优化算法及其应用研究　出处：《广西民族大学》2016年硕士论文　论文类型：学位论文

【摘要】：风力驱动优化算法是一种新的启发式算法,是由Z.Bayraktar等人在2010年提出。风力驱动优化算法的思想是模拟自然现象由于各地区气压不同而导致空气流动,最终达到气压平衡的过程。在大气中,空气的流动是为了尝试去平衡气压。当气压失衡时,空气将受到气压梯度力,继而以一定的速度从高压地区向低压地区流动,最终使气压达到平衡。由于风力驱动优化算法具有结构简单、控制变量少、易于理解和实现等优点,算法自提出以来受到了越来越多的学者关注。但该算法也存在着前期收敛速度过快,易陷入局部最优,后期种群多样性不足导致收敛速度慢等缺陷,极大地限制了风力驱动优化算法的应用范围。因此,风力驱动优化算法无论是在理论方面,还是在应用方面,都有待于进一步的研究和扩展。本文针对风力驱动优化算法存在的不足进行分析,从更新策略和编码方法等方面对算法进行改进,并将改进后的算法应用到实际优化问题中。工作内容主要包括三个方面:(1)采取双种群策略对风力驱动优化算法进行改进,其中,一个种群由风力驱动优化算法进行更新,另一个种群由差分进化算法进行更新,两个算法通过信息共享机制实现种群的共同进化。该策略能够增加种群多样性,继而增强算法的全局搜索能力,避免算法因收敛速度过快而陷入局部最优。(2)通过改变个体的编码方式来改进风力驱动优化算法的性能。将复数编码的思想应用到风力驱动优化算法中,利用实部和虚部两个变量来表示一个自变量。由于每个复数都可以表达二维信息,这样能增强种群表示的信息量和个体的多样性。在复数编码风力驱动优化算法的基础上引入贪心策略,有效地解决0-1背包问题。(3)将量子编码理论的思想应用到风力驱动优化算法中,提出了一种量子风力驱动优化算法。利用量子旋转门实现种群的更新,利用量子非门策略实现种群个体的变异。这两个策略能够提高种群的多样性,避免种群过早收敛。将量子风力驱动优化算法应用于无人机航路规划,表明了算法的有效性及可行性。
[Abstract]:Wind driven optimization algorithm is a new heuristic algorithm. In 2010, Z. Bayraktar et al., the idea of wind driven optimization algorithm is to simulate the air flow caused by the different air pressure in different regions. The process of eventually reaching a barometric equilibrium. In the atmosphere, air flows in an attempt to balance the pressure. When the pressure is out of balance, the air is subjected to a pressure gradient. Then flow from high pressure region to low pressure area at a certain speed, and finally make the pressure balance. Because of the advantages of simple structure, less control variables, easy to understand and realize, wind driven optimization algorithm has the advantages of simple structure, less control variables, and easy to understand and realize. Since the algorithm was put forward, more and more scholars have paid attention to it. However, the algorithm also has some shortcomings such as the early convergence speed is too fast, it is easy to fall into local optimum, and the late population diversity is insufficient, resulting in the slow convergence rate and so on. The application of wind driven optimization algorithm is greatly limited. Therefore, wind driven optimization algorithm is not only in theory but also in application. This paper analyzes the shortcomings of the wind-driven optimization algorithm and improves the algorithm from the aspects of update strategy and coding method. And the improved algorithm is applied to the practical optimization problem. The main work includes three aspects: 1) adopting the dual-population strategy to improve the wind-driven optimization algorithm. One population is updated by wind driven optimization algorithm and the other population is updated by differential evolution algorithm. The two algorithms realize the coevolution of population through information sharing mechanism. This strategy can increase population diversity. Then the global search ability of the algorithm is enhanced. To avoid falling into local optimum because of the fast convergence speed, the algorithm can improve the performance of wind driven optimization algorithm by changing individual coding method. The idea of complex number coding is applied to wind driven optimization algorithm. Two variables, real part and imaginary part, are used to represent an independent variable. In this way, the amount of information and the diversity of individuals can be enhanced, and the greedy strategy is introduced on the basis of the complex coded wind-driven optimization algorithm. Effectively solve the 0-1 knapsack problem. (3) the quantum coding theory is applied to the wind driven optimization algorithm, and a quantum wind driven optimization algorithm is proposed, which uses the quantum rotary gate to update the population. The two strategies can improve population diversity and avoid premature convergence. Quantum wind driven optimization algorithm is applied to UAV route planning. The effectiveness and feasibility of the algorithm are demonstrated.
【学位授予单位】：广西民族大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP18

【参考文献】