群体智能优化算法

A Survey of Swarm Intelligence Optimization Algorithm

王辉 钱锋

讨论四种群体智能优化算法——蚁群算法、微粒群算法、人工鱼群算法和混合蛙跳算法,对其算法的原理、发展及应用进行了综述。提出了群体智能优化算法统一框架模式,并对群体智能优化算法进一步发展进行了讨论。[著者文摘]

　Discussed four swarm intelligence optimization algorithms：ant colony optimization algorithm（ACO）,particle swarm optimization algorithm（PSO）,artificial fish school algorithm（AFSA） and shuffled frog leaping algorithm（SFLA）.The principle,development and applications of them were summarized.Put forward unified frame pattern of swarm intelligent optimization algorithm and discussed the future direction and content of it.[著者文摘]

欧氏Steiner最小树问题的智能优化算法

Intelligent Optimization Algorithms for Euclidean Steiner Minimum Tree Problem

金慧敏 马良 王周缅

欧氏平面内连接固定原点的最小树长问题，即欧氏Steiner最小树问题，为组合优化中的NP难题，因此合理的方法是寻找启发式算法。该文给出了两种智能优化算法——模拟退火法和蚂蚁算法。首先概述智能优化算法并将中面划分成网格，然后分别介绍两种算法的原理及实现过程，最后通过一系列计算实验，测试了算法的运行性能，获得了较好的效果。[著者文摘]

The Euclidean Steiner tree problem, which concerns the construction of a tree that connects a given set of terminal points in Euclidean space with the minimal total lengths, is NP-hard in eomhinatorial optimization. This paper presents two intelligent optimization algorithms-the simulated annealing algorithm and the ant algorithm. Firstly, this paper summarizes the intelligent optimization algorithm and divides space into grids, then introduces the principles of two algorithms and their implementations. Finally, it implements serics of computational experiments and tests the performances, that result within satisfaction.[著者文摘]

多智能体遗传算法用于马斯京根模型参数估计

Application of multi-agent genetic algorithm to parameter estimation of Muskingum Model

鲁帆 蒋云钟 王浩 牛存稳

将智能体对环境的感知和反作用的能力与遗传算法的搜索方式相结合提出了一种改进的多智能体遗传算法，用于马斯京根模型的参数估计。该方法中每个智能体代表一个候选解并固定在网格上，为了增加自身能量，它将与其邻域的智能体进行合作或竞争，也可以利用自身的知识进行自学习来增加能量，通过这些智能体与智能体间的相互作用来达到优化模型中参数的目的。应用实例表明，该算法同其他算法相比具有更好的优化性能，从而为准确估计马斯京根模型参数提供了一种更为有效的方法。[著者文摘]

The ability of apperception and counteractive to environment of agent is combined with searching mode of genetic algorithm to establish an improved multi-agent algorithm for estimating the parameter of Muskingum Model. In this model every agent represents an optional solution and is fixed in the grid. In order to elevate its self energy every agent competes or cooperates with its neighbors, and it can also use its knowledge to study for elevating the energy. Through the interaction between the agents the objective of optimizing the parameters can be realized. The advantage of this proposed method is demonstrated by an example.[著者文摘]

一种基于RLS和RVSSLMS智能天线波束赋形算法

Beamforming Algorithm Based on RLS and RVSSLMS for Smart Antenna

赵鹏飞 刘刚

根据智能天线波束赋形算法必须考虑设备的复杂性和收敛速度的要求,提出一种结合RLS和RVSSLMS算法各自优点的RLS-RVSSLMS波束赋形算法,并用Matlab进行了仿真。仿真结果表明：RLS-RVSSLMS既具有RLS算法收敛速度快的特点,同时保持了LMS算法计算量小的特点。[著者文摘]

According to the condition that the beamforming algorithm must take the complexity of equipment and convergence speed into account, a beamforming algorithm which combines the advantages of RLS algorithm and RVSSLMS algorithm respectively is presented. Simulating it using Matlab, the results indicate that the convergence speed is faster as RLS and the operation is less as RVSSLMS in RLS-RVSSLMS algorithm.[著者文摘]