首页 | 本学科首页   官方微博 | 高级检索  
相似文献
 共查询到20条相似文献,搜索用时 281 毫秒
1.
贝叶斯优化算法是利用贝叶斯网络匹配进化种群的优良解集而产生新的染色体来体现种群的进化.在贝叶斯网络对种群进行匹配的过程中,贝叶斯网络结构越复杂,种群的进化信息描述越完整,进化质量越高,但运算速度相对来说越慢;相反,贝叶斯网络越简单,算法描述的种群的进化信息越少,进化质量越差,但却能够提高算法的运算速度.基于此,给出了简单贝叶斯优化与复杂贝叶斯优化定义.针对简单贝叶斯网络提出了基于BD度量的三步结构学习算法,并给出了一个利用这种算法进行贝叶斯网络结构学习的例子.  相似文献   

2.
在大规模高维优化问题中,随着决策变量数目的增加,协同进化算法在搜索全局最优解过程中容易陷入局部最优。基于此,提出了一种基于协方差分析的合作协同进化差分进化算法,在根据决策变量之间的相关性对优化问题进行分组之后,针对子组件内部变量之间的相关性会影响种群进化过程的现象,在对子组件优化的过程中,利用协方差计算种群分布的特征向量,通过坐标旋转消除变量之间的相关性,有效避免在种群搜索过程中陷入局部最优,同时加快了算法的寻优速度。在CEC2014测试函数集上进行了对比实验,实验结果表明,所提算法具有可行性。  相似文献   

3.
万寿红  梁肖  岳丽华  熊焰 《电子学报》2011,39(2):375-377
 本文在快速进化规划基础上,提出了一种解决组合优化问题的自适应柯西进化规划ACEP.该算法融合了柯西变异的优点,通过调整参量r来适当的改变搜索的步长,相对于经典进化规划CEP和快速进化规划FEP只需一半的种群数量便可快速到达问题的最优解,最后0/1背包问题的对比实验结果表明了其优越性.  相似文献   

4.
基于信息熵差异性度量的数据流增量集成分类算法   总被引:2,自引:0,他引:2  
琚春华  邹江波 《电信科学》2015,31(2):92-102
对分类器之间的差异性进行了研究,提出了一种基于信息熵差异性度量的增量集成分类算法,将信息熵差异性度量方法融入到基分类器选择过程中,通过对训练数据集的基分类结果的信息熵差异度计算,采用循环迭代优化的选择方法,以熵差异性最优化为约束目标,动态调整基分类器个数,实现了分类准确稳定,减少了系统开销。通过实验比对,证明了算法在数据流处理时比其他算法具有更小的开销和较强的适应性。  相似文献   

5.
基于改进免疫遗传算法的带硬时间窗车辆调度问题的实现   总被引:2,自引:0,他引:2  
免疫算法是模仿生物体高度进化,复杂的免疫系统仿生的一种智能化启发式算法。带硬时间窗的车辆路径问题(VSPHTW)是在基本的车辆路径问题(VSP)上增加了时间窗约束条件的一种变化形式,是一个典型的NP难题。通过采用一种改进的信息熵计算方法、交叉和变异概率的自适应机制,构造一个改进的免疫算法来求解VSPHTW,并将求解结果与其他遗传算法比较。比较结果显示,该算法对于求解VSPHTW问题具有较好的性能。  相似文献   

6.
高维数据中存在着成千上万个特征,大量的特征导致问题搜索空间过大,增加了计算代价,影响了数据分类预测的准确性.为了提高特征选择的效率,本文提出了一种对称不确定性和种群降维机制的粒子群特征选择算法,该算法设计了一种基于对称不确定性指标的初始化方法,降低特征选择的计算代价.通过非支配排序的种群降维机制,减少进化过程中冗余特征的影响.在5个公开生物医学的高维数据集上的实验结果表明,该算法能够针对高维数据特征选择问题取得更好的分类精度和更小的最优子集特征个数,并在时间运行方面有一定的优势.  相似文献   

7.
探讨了元胞遗传算法中种群多样性对全局寻优/局部收敛平衡的意义,提出了基于邻域结构内元胞遗传算法的多样性度量方式,并提出了改变遗传算子的元胞遗传算法来维持进化过程种群的多样性,算法将元胞空间网格嵌入到种群空间中,模拟遗传操作在相邻个体之间进行。该算法不仅提高了全局搜索能力,且在维持种群多样性方面有一定优势。  相似文献   

8.
受信息分布特征和测量技术的限制,隐式信息场具有显著的观测局限性,通常观测者无法获得未到达区域的数据,从而导致寻源行为具有强烈的不确定性。受生物寻源行为的启发,文中从搜索行为与信息获取和利用的角度出发,提出了一种基于动态均衡策略的自主寻源方法。首先,借助多目标优化理论,建立寻源数学模型,将隐式信息场寻源任务归结为多目标收敛问题;其次,借鉴进化算法的思想,将运动搜索与种群进化相结合,构建以可行性搜索行为为个体样本的进化种群,通过对多目标问题的求解,引导载体完成寻源任务;然后,在寻源过程中,引入分布熵度量搜索偏向,结合寻源进程的需求,设计了一种动态均衡搜索的寻源方法,并给出寻源算法;最后,通过理论分析和仿真验证,证实了所提方法的有效性和合理性。  相似文献   

9.
量子克隆进化算法   总被引:1,自引:1,他引:0  
刘芳  李阳阳 《电子学报》2004,31(B12):2066-2070
本文在量子进化算法的基础上结合基于克隆选择学说的克隆算子,提出了改进的进化算法——量子克降进化策略算法(QCES).它既借鉴了量子进化算法的高效并行性又利用克隆算子来代替其中的变异和选择操作,以增加种群的多样性,避免了早熟,且收敛速度快.本文不仅从理论上证明了该算法的收敛,而且通过仿真实验表明了此算法的优越性.  相似文献   

10.
贾天理  梅杨 《信息技术》2011,(2):110-112,128
基于信息熵概念与算法模型,探讨了物联网信息度量的科学问题,提供了信息有效度量的方法,分析了信息熵的物联意义,最后描述了物联网的广泛应用场景。  相似文献   

11.
为了降低多元低密度奇偶校验(Low-density parity check,LDPC)码Min-max译码算法的运算量,提出一种自适应Min-max(Adaptive Min-max,AMM)译码算法.该方法以Min-max算法为基础,以每次迭代后的校验节点错误率(Check-node Error Rate,CER)为调节参数,采用自适应算法对变量节点的向量长度进行截短,去除置信度较低的分量,仅对置信度较高的分量进行更新.当CER降低到一定程度时,对校验节点个数进行自适应截短,仅对不满足校验方程的校验节点进行消息迭代更新,进一步降低AMM算法的复杂度.仿真结果表明,在相同误码性能条件下,AMM算法运算量较固定长度截短的Min-max算法减少20%.  相似文献   

12.
The segmentation effect of Tsallis entropy method is superior to that of Shannon entropy method, and the computation speed of two-dimensional Shannon cross entropy method can be further improved by optimization. The existing two-dimensional Tsallis cross entropy method is not the strict two-dimensional extension. Thus two new methods of image thresholding using two-dimensional Tsallis cross entropy based on either Chaotic Particle Swarm Optimization (CPSO) or decomposition are proposed. The former uses CPSO...  相似文献   

13.
In downlink multiuser multiple-input multiple-output (MU-MIMO) systems, block diagonalization (BD) is a well-kown precoding technique that eliminates interuser interference. The number of simultaneously supportable users with BD is limited by the number of base station transmit antennas and the number of user receive antennas. The brute-force search for the optimal user set, however, is computationally prohibitive. Therefore, we propose a low complexity and suboptimal user selection algorithm based on block diagonalization for MU-MIMO systems. We introduce a strong tight upper bound of sum capacity as selection metric. Furthermore, we employ a substitution operation to improve system performance. The computational complexity analysis and simulation results show that the proposed algorithm achieves comparable throughput with low complexity compared to the existing algorithms.  相似文献   

14.
In this paper, we address a user scheduling (selection) problem in the uplink multiuser multiple input multiple output (MIMO) wireless communication system. For this problem, the computational complexity of exhaustive search grows exponentially with the number of users. We present an iterative, low-complexity, sub-optimal algorithm for this problem. We apply an Estimation of Distribution Algorithm (EDA) for the user scheduling problem. An EDA is an evolutionary algorithm and updates its chosen population at each iteration on the basis of the probability distribution learned from the population of superior candidate solutions chosen at the previous iterations. The proposed EDA has a low computational complexity and can find a nearly optimal solution in real time for the user scheduling problem. Beyond applying the general EDA to user scheduling, we also present specific improvements that reduce computation for obtaining an acceptable solution. These improvements include the idea of generating an initial population by cyclically shifting a candidate solution. The simulation results show that our proposed algorithm performs better than other scheduling algorithms with comparable complexity.  相似文献   

15.
肖锋  周杰 《电子学报》2013,41(4):757-762
切平面法作为求解非光滑凸优化问题的典型方法,在支持向量机问题的求解中得到了广泛的应用.但是该算法在求解过程中往往会出现不稳定的情况.针对这一不稳定性,前人提出了优化切平面法,通过在切平面法中加入线搜索环节来确保目标函数单调下降.但是优化切平面法的运算复杂度比较高,不适合训练数据量大、对训练速度要求高的应用.本文提出了一种基于活跃集的优化切平面法,在计算目标函数和进行线搜索时,只单独处理活跃集内的样本,将其它样本当作一个整体来进行处理.相对于传统的优化切平面法,本文方法只需在一部分样本上计算目标函数和进行线搜索,从而可以在不损失求解精度的前提下节省求解时间.  相似文献   

16.

This article proposes an improved Newton algorithm as a low complexity signal detection scheme for linear receiver in large scale multiple- input multiple- output (LS-MIMO) single carrier frequency division multiple access (SC-FDMA) uplink system, where a large number of antennas are set up at the base station and active users are with a single antenna system. Data detection for uplink SC-FDMA system is one of the specific challenges due to the significant rise in the dimension of antennas and number of subcarriers. Especially for symbol detection process, LS-MIMO SC-FDMA system with linear detector requires to perform a large matrix inverse computation. Even though linear detectors such as zero forcing (ZF) and minimum mean square error (MMSE) can achieve near-optimal detection performance, they still introduce high computational complexity and obliviously involve in the computation of matrix inversion. Therefore, a design of complexity reduction algorithm based near-optimal detector for LS-MIMO SC-FDMA system attains research interest. The improved Newton algorithm is employed to obtain linear detection solution which iteratively performs matrix free-inversion operation. The new algorithm performs matrix–matrix multiplication into matrix–vector multiplication, which substantially reduces receiver detection complexity. The efficacy of the proposed method is investigated at 16-QAM. Both ZF and MMSE criteria are proposed and compared through simulations. Simulation results illustrate that the proposed scheme outperforms the conventional detection schemes and exhibits near-optimal performance with a small number of iterations. Further, bit-error-rate performance is closer to classical linear detector with affordable computational complexity.

  相似文献   

17.
We present a graph theoretical methodology that reduces the implementation complexity of the multiplication of a constant vector and a scalar. The complexity of implementation is defined as the required amount of computations like additions. The proposed approach is called minimally redundant parallel (MRP) optimization and is mainly presented in a finite impulse response (FIR) filtering framework to obtain a low-complexity multiplierless implementation. The key idea is to expand the design space using shift inclusive differential coefficients (SIDCs) together with computation reordering using a graph theoretic approach to obtain maximal computation sharing. The problem is formulated using a graph in which vertices and edges represent coefficients and computational cost (number of resources). The multiplierless solution is obtained by solving a set cover problem on the vertices in the graph. A simple polynomial run time algorithm based on a greedy approach is presented. The proposed approach is compared with common-subexpression elimination to show slightly better results and is combined with it for further reduction of complexity. Simulation results show that 50-60% complexity reduction is achieved by only applying the MRP algorithm, and 70% complexity reduction is obtainable by combining it with common-subexpression elimination under a delay constraint of two or three.  相似文献   

18.
基于LWE的集合相交和相等的两方保密计算   总被引:2,自引:0,他引:2  
利用格上LWE(Leaning With Error)困难性假设,将保密地比较两个数是否相等转化为判断对随机串加密后的解密是否正确,有效地解决了数和集合关系的判定、求集合交集和集合相等安全多方计算问题,并利用模拟范例证明该协议在半诚实模型下是安全的。与传统的基于数论的协议相比,该方案由于不需要模指数运算,因而具有较低的计算复杂度,同时因其基于格中困难问题,因而能抵抗量子攻击。  相似文献   

19.
随机最大似然算法(Stochastic Maximum Likelihood,SML)具有优越的波达方位(Direction-of-Arrival,DOA)估计性能,但SML解析过程较高的计算复杂度限制了该算法在实际系统中的应用.针对SML计算复杂度高的问题,提出一种低复杂度的粒子群优化算法(Particle Swarm Optimization,PSO),解决了传统PSO算法中粒子数多和迭代次数多的双重缺点.首先,根据天线获得的信号,将旋转不变子空间法(Estimation of Signal Parameters via Rotational Invariance Techniques,ESPRIT)求得的闭式解作为DOA的预估计值,同时计算系统此时的信噪比以及SML在此信噪比下的克拉-美罗界(Cramer-Rao bound,CRB).然后,根据DOA预估计值和当前CRB值在SML最优解的近邻范围内确定较小的初始化空间,并在该空间初始化少量粒子.最后通过设计合适的惯性因子w,使粒子以合理的速度搜索最优解.实验结果表明,改进PSO算法所需的粒子个数和迭代次数大约是传统PSO算法的1/5,降低了SML的解析复杂度,计算时间是传统PSO算法的1/10,因此在收敛速度上也有显著的优势.  相似文献   

20.
针对新一代视频压缩编码标准HEVC计算复杂度较高的特点,利用视频序列间时域上的相关性,提出了一种基于灰度差值的编码单元快速划分策略.该策略根据当前编码块与参考块之间的灰度差值进行运动条件判决,在进行编码之前提前确定当前编码单元的编码深度信息,减少帧间预测编码的次数,从而有效地降低了编码端的计算复杂度.实验结果表明,该算法在编码效率和峰值信噪比(PSNR)损失都很小的情况下,和HM标准中的帧间预测算法相比,平均降低了50.18%的编码时间.  相似文献   

设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号