【作者】 马骏；

【导师】 张健沛；

【作者基本信息】 哈尔滨工程大学 ， 计算机应用技术， 2009， 博士

【摘要】 随着Internet的迅速发展,网络技术和分布式人工智能领域不断取得新的突破,传统的分布式计算模式已经不能满足当前异质网络上复杂的分布式计算要求,移动agent技术就是在这种情况下诞生的一种新的网络计算技术。移动agent具有自主性、协作性以及移动性,它可以代替用户去完成所需的任务,而且它能根据用户各种需求以及实际情况在网络中自主地迁移。这种模式非常适合分布式计算的要求,为分布式计算环境中合理、有效的组织信息,以及信息的访问和共享提供了新思路和新方案。移动agent所支持的计算模式克服了传统信息管理和共享方式的弊端,提高了分布式环境下信息共享和获取能力。移动agent最大的特点是其移动性,合理的迁移路径和迁移策略将使得移动agent的性能得到极大的改善。旅行agent问题是一种复杂的组合优化问题,目的在于解决移动agent在不同主机间移动时,如何搜索最优的迁移路径,它是移动agent迁移路径规划中最经典的问题。针对蚁群算法在解决此类问题时需要较长搜索时间和易于陷入局部极小等缺点,引入变异运算,并且对蚁群算法的全局和局部更新规则进行了改进,降低了蚁群算法求解旅行agent问题时,陷入局部极小而导致系统出现停滞现象的可能。仿真实验结果表明,改进的蚁群算法使得移动agent能够以更优的效率和更短的时间来完成任务。针对遗传算法、粒子群优化算法求解移动agent路径选择问题时,进入后期搜索效率较低等问题。对移动agent的路径选择问题进行形式化描述,给出该问题的多约束最优路径求解模型,根据模拟退火算法中的Metropolis准则接受最优单体以推动文化算法中信念空间的进化,提出了一种将模拟退火算法嵌入文化算法框架中来求解移动agent迁移路径选择的方法。仿真实验结果表明,应用文化算法求解移动agent迁移路径选择问题,具有较好的结果以及较低的运算代价。针对现有的移动agent迁移策略在迁移过程中不能为移动agent动态选择迁移主机的问题。结合支持向量机的相关原理,在已有移动agent规范和模型的基础上,改进了移动agent的模型,提出了基于支持向量机的移动agent迁移策略,并给出了相应的智能迁移模型。基于支持向量机的移动agent迁移策略,充分利用了支持向量机优异的学习能力以及分类性能,力图为移动agent规划出最佳迁移目的主机,缩短移动agent完成任务的时间。仿真实验结果表明,与一步迁移策略以及理想迁移策略相比,该策略能够以较大的概率获得最优解。移动agent的迁移安全是移动agent技术现阶段面临的主要安全问题之一。针对Domingo所提出的移动agent迁移协议不能防止恶意主机联合篡改移动agent迁移信息的安全问题。在此协议的基础上,利用hash函数,提出了一个基于Merkle树的移动agent动态安全迁移协议,并对其安全性和计算复杂度进行了详细的分析。相比已有的方案,基于Merkle树的移动agent动态安全迁移迁移协议在保证安全性的同时,计算复杂度得到了明显的降低。更多还原

【Abstract】 With the development of Internet, network technology and distributed artificial intelligence field have continuously achieved new breakthrough. The traditional distributed computing model can not meet the demands of complex distributed computing on heterogeneous network. Based on theses studies, mobile agent technology which is a new network computing technology is proposed. Mobile agent has the character of autonomy, collaboration and mobility. Mobile agent can independently move in heterogeneous network and seek the appropriate computation resources. It will migrate from machine to machine and accomplish the specific task on behalf of the user. The model is fit for the demand of distributed computing. Mobile agent technology provides new thoughts and methods for the information organization, information high efficiency access and information sharing in distributed computing environment. The computing model of mobile agent can overcome the disadvantages of traditional information management and information sharing and improve the information sharing and acquisition capacity in distributed computing environment. The remarkable characteristic of mobile agent is mobility. Reasonable routing policy and routing planning will obviously improve the system performance of mobile agent.Traveling agent problem is a complex combinatorial optimization problem, which solves the problem of planning out an optimal migration path when agents migrate to several hosts. In this paper, an improved ant colony algorithm is presented. A mutation operator is introduced and the local and global updating rules of pheromone are modified on the basis of ant colony algorithm. The algorithm greatly decreases the possibility of halting the ant system due to arriving at local minimum. The simulation experiment results show that mobile agent can accomplish the computing task with high efficiency and short time.Aiming at genetic algorithm and particle swarm optimization has lower searching efficiency in solving route choice of mobile agent. The key idea behind cultural algorithm is to explicitly acquire problem-solving knowledge from the evolving population and in return apply that knowledge to guide the search. In this paper, routing problem of mobile agent is formally demonstrated; also solving model of Multi-Constrained non-dominated optimal route is presented. Cultural algorithm is designed to solve the problem of mobile agent’s routing: accepting the best individuals to improve the evolution of belief space by simulated annealing, search step length as situational knowledge is used to guide searching of the optimal solution in population space. The simulation experiment shows that the algorithm produces highly competitive results at a relatively low computational cost.Aiming at the existed mobile agent migration strategy can not dynamically choose migration host for mobile agent. Based on the existing model and criterion of mobile agent, machine learning theory is combined with the correlation theory of support vector machine and the model of mobile agent is improved. Intelligent migration strategy and model of mobile agent based on support vector machine are proposed. Mobile agent with support vector machine can perceive the changes of environment, react immediately, embodying the reactivity and autonomy of agent. Compared with other migration strategies, it can obtain the optimal result with high probability. The simulation experiment results on Aglet platform show that the migration strategy is effective and available.Migration security is one of main security problem of mobile agent. In this paper, we analyze the existing effective migration protocol and point the protocol has serious security hidden trouble:it is not against collusion of malicious hosts. Based on this protocol, using hash function, a security itinerary protection of mobile agent based on Merkle trees is proposed. Security and computational complexity are discussed in detail. According to the existing protocol, improved protocol meets the demand of security. Computational complexity is reduced.更多还原