节点文献
基于信息熵遗传算法的舰船导航路径规划技术研究
Research on Path Planning of Warship Navigation Based on Information Entropy-Genetic Algorithm
【作者】 李魁星;
【导师】 钱华明;
【作者基本信息】 哈尔滨工程大学 , 制导与控制, 2010, 硕士
【摘要】 随着科技的不断进步,舰船导航正逐渐向智能化的方向发展。其中路径规划是舰船智能航行的关键技术之一。本文以舰船路径规划为背景,以电子海图系统为平台,对舰船在真实海洋环境下的路径规划问题进行了深入的研究。将信息熵理论与遗传算法相结合,并根据路径规划技术特点,提出了应用于路径规划领域的信息熵遗传算法,克服了传统遗传算法在路径规划时耗时长、易陷入局部解的缺点。针对电子海图显示特点,本文采用多边形障碍表达方式和路径点编码方式,并将遗传算法作为基本算法;为描述种群多样性,重点研究了信息熵原理,提出了基于栅格思想的路径种群熵概念,实现了种群多样性的测量;在此基础上,进一步提出了基于路径种群熵的双轮盘赌选择策略,给出了基于路径种群熵的自适应交叉、变异概率计算公式,达到了在算法前期保持种群多样性,而在算法后期加快收敛速度的目的,提高了算法性能,克服了算法缺陷;提出了先启发生成再优化的初始种群生成方法,提高了初始种群的个体质量;提出了基于区域的碰撞检测方法,减少了检测次数,提高了算法效率。最后,利用本文设计算法进行了路径规划仿真,得到了理想的结果,验证了本文算法的有效性。同时与原有的自适应遗传算法进行了仿真对比,通过实际效果的差异验证了本文算法的优越性。
【Abstract】 As technology progresses, the ship navigation are gradually moving towards the direction of the development of intelligent and path planning is one of the key technologies of it. In the background of path planning of ship and on the platform of electronic chart, the paper makes the thorough research in path planning for ship sailing in the real marine environment. Information entropy-genetic algorithm which applied to the field of path planning is proposed based on the integration of information entropy and genetic arithmetic and the characteristics of path planning. The defects of traditional genetic algorithm in path planning, time-consuming and falling into local solutions are overcome.For the characteristics of the display of electronic chart, this paper adopts polygonal obstacles and path-point coding presentation, and genetic algorithm as the basic algorithm. To describe the diversity of population, a concept of path population entropy is proposed based on grid concept and focusing research on the information entropy principle and realize the diversity of population measurement. On this basis, the paper further proposed the dual -wheeled-gambling selection strategy based on path population entropy and given the adaptive formula of variation rate and cross rate based on path population entropy, reaches purpose of maintaining diversity of population in the early part of algorithm and speeding up the convergence of the algorithm in the latter part of algorithm. A method of createing initial population which optimizeing path after creating path is proposed and the quality of the initial population is improved. To reduce the detection times, a region-based collision detection method is proposed and the efficiency of the algorithm is improved.Finally, the use of algorithms designed in this paper a path planning simulation, has been the ideal outcome, validate the effectiveness of the algorithm. At the same time with the original self-adaptive genetic algorithm simulation comparison, through the difference between the actual results validate the superiority of the algorithm.
【Key words】 Path planning; Genetic algorithm; Information entropy; Path population entropy;