【作者】 张频捷；

【导师】 谭冠政；

【作者基本信息】 中南大学 ， 控制科学与工程， 2010， 硕士

【摘要】 随着人们对生命本质的研究不断深入,生命科学正以空前的速度发展,人工智能的研究也开始摆脱传统逻辑计算之束缚,大胆探索起新的非经典计算途径。人工智能先驱Minsky曾认为：“我们应该从生物学而不是物理学受到启示……”。在这种背景下,蚁群算法应运而生,1992年,意大利学者M.Dorigo通过研究蚁群的觅食行为,在他的博士论文中提出了一种基于种群的模拟进化算法——蚁群优化(ACO)。由于蚁群算法具有稳健性、全局性、普遍性、分布式计算等众多优点,之后引起了学者们的极大关注,与其它较为成熟算法的相比较,蚁群算法在解决许多实际问题时有着更加优越的表现,在过去十多年的时间里,蚁群算法发展迅速,已在组合优化、网络路由、函数优化、数据挖掘、机器人路径规划甚至股票投资等领域获得了广泛的应用,并取得了较好的成果。本文围绕蚁群算法的理论及其应用,首先详细阐述了算法的生物学机理,介绍了基本蚁群算法的原理、模型、特点及其在TSP中的实现,并对蚁群算法在各领域的应用做了简要的叙述。同时,指出了基本蚁群算法的一些不足,然后针对不足之处,列举了一些典型的改进算法,例如ACS,MMAS等,并通过实验以及其他学者的研究成果说明了基本蚁群算法模型中各参数的合理选择方法。其次,本文简要介绍了多机器人系统,指出多机器人系统的优势与研究的必要性,再根据ACS算法的思想,结合多机器人的协作,将其应用于多机器人路径规划之中。最后,本文利用蚁群算法和FCM算法的混合算法,结合边缘检测技术,将其应用于彩色图像分割之中,并以几种不同的代表性图片为样本,进行了计算机仿真,实验取得了很好的结果,表明此混合算法是一种较好的彩色图像分割方法。更多还原

【Abstract】 As the further studies of the essence of life,life sciences is develop at an unprecedented pace,Artificial intelligence research have begun to break away from the bondage of traditional Logical computation, and explore the new non-classical computation approach.The pioneer of Artificial intelligence Minsky had thought:"We should be enlightened from biology rather than physics."In background of this, ant colony algorithm emerged,in the year 1992, the Italian scholar M.Dorigo by studying the ant colony’s foraging behavior, in his doctoral thesis propose a simulation evolutionary algorithm based on Colony——ant colony optimization (ACO).As the ant colony algorithm is steady, holistic, universal, distributed computing and many other advantages, attract a great attention of many scholars.compared with other more mature algorithm, ant colony algorithm has a more superior performance to solve some practical problems,In the past ten years,the development of ant colony algorithm is very fast.In combinatorial optimization, network routing, function optimization, data mining, robot path planning and even in the area of stock investment ant colony algorithm has been widely applied,and achieved good results.This paper focuses on the theory and application of ant colony algorithm, first, we elaborate on the biological mechanism of the algorithm, introduced the theory of basic ant colony algorithm, model, characteristics and the realization in TSP, and give a brief description about the application of ant colony algorithm in various fields.At the same time, pointed out some deficiencies about the basic ant colony algorithm, and then for the deficiencies we cited a number of typical improvement algorithms, such as ACS,MMAS,etc.;Then we through experiments and other scholars’s research explain the parameter selection method in the basic ant colony algorithm models.Then this paper introduces the multi-robot system, and point out that the advantages of multi-robot systems and the necessity of the research.According to the theory of ACS,combination with multi-robot collaboration, this algorithm was applied to multi-robot path planning.Finally, we use the synthetical algorithm of ant colony algorithm and FCM algorithm, combination of edge detection techniques,and apply it to the colorized image segmentation, some typical images were used as experimental samples, we make a computer simulation, and the experiment achieved excellent results.It shows that this synthetical algorithm is an excellent approach for colorized image segmentation.更多还原