推广模拟退火方法及其应用

本文介绍了推广模拟退火方法的基本思想及其统计基础。我们通过一系列标准函数测试了推广模拟退火方法的性能。作为该方法的应用 ,我们研究了Thomson问题和Ni团簇结构。最后 ,探讨了推广模拟退火方法的效率随体系复杂性的变化。     

薄膜椭偏测量的自适应混合反演算法

基于梯度的线性反演方法计算效率高,基于随机扰动的模拟退火方法寻找最优解能力强针对薄膜椭偏测量的多极值问题,综合两者的优点,提出一种求解薄膜椭偏测量问题的混合反演算法.模型每次扰动采用线性寻优方法搜寻局部最优解,叠代过程中采用均匀设计的模拟退火方法随机搜寻模型,使该算法有跳出局部最优解的能力,可以在较少的叠代次数内搜寻到全局最优解,从而提高求解薄膜椭偏测量非线性反演方法的计算效率.对反演过程控制参数进行讨论,该算法具有自适应的特点.计算表明,该算法可有效求解薄膜椭偏测量的多极值问题.     

Influence of rounding errors on the quality of heuristic optimization algorithms

Search space smoothing and related heuristic optimization algorithms provide an alternative approach to simulated annealing and its variants: while simulated annealing traverses barriers in the energy landscape at finite temperatures, search space smoothing intends to remove these barriers, so that a greedy algorithm is sufficient to find the global minimum. Several formulas for smoothing the energy landscape have already been applied, one of them making use of the finite numerical precision on a computer. In this paper, we thoroughly investigate the effect of finite numerical accuracy on the quality of results achieved with heuristic optimization algorithms. We present computational results for the traveling salesman problem.     

Optimization of a total internal reflection lens by using a hybrid Taguchi-simulated annealing algorithm

In this paper, we propose a new method for optimization of a total internal reflection (TIR) lens by using a hybrid Taguchi-simulated annealing algorithm. The conventional simulated annealing (SA) algorithm is a method for solving global optimization problems and has also been used in non-imaging systems in recent years. However, the success of SA depends heavily on the annealing schedule and initial parameter setting. In this study, we successfully incorporated the Taguchi method into the SA algorithm. The new hybrid Taguchi-simulated annealing algorithm provides more precise search results and has lower initial parameter dependence.     

玉米秸秆纤维素和半纤维素NIRS特征波长优选

预处理是提高玉米秸秆生物转化利用效率的有效途径。玉米秸秆经生物炼制转化为生物燃料时,转化率与其原料内的纤维素和半纤维素含量直接相关。为了实现对预处理后玉米秸秆的生物炼制过程的有效调控,提出使用近红外光谱(NIRS)对玉米秸秆的纤维素和半纤维素含量进行快速检测,解决传统化学方法测试速度慢、成本高的问题。为了提高NIRS检测的效率和精度,将遗传算法与模拟退火算法相结合构建遗传模拟退火算法(GSA)用于预处理后玉米秸秆纤维素和半纤维素含量NIRS特征波长优选。GSA算法以NIRS波长点数为码长进行二进制编码,以偏最小二乘法(PLS)回归模型的交叉验证均方根误差为目标函数,结合温度参数设计适应度函数,基于Metropolis判别准则实现扰动解的选择复制,能够在避免早熟的同时有效提高进化后期的搜索效率。采用碱预处理、生物预处理及其相结合的方法对采集的玉米秸秆进行预处理后制备样品120个,并测定其纤维素和半纤维素含量及NIRS。使用7点Savitzky-Golay平滑结合多元散射校正和标准正则变换对光谱进行预处理后,利用Kennard-Stone法按3∶1比例划分校正集和验证集。然后,使用GSA算法对NIRS全谱进行特征波长优选(记为Full-GSA)、对协同区间偏最小二乘法(SiPLS)优选后谱区进行特征波长优选(记为SiPLS-GSA)、对反向区间偏最小二乘法(BiPLS)优选后谱区进行特征波长优选(记为BiPLS-GSA),并使用PLS回归模型和验证集对特征波长优选结果进行评测。Full-GSA以全谱1 557个波长点为基因,执行16次算法,优选出118个纤维素特征波长点和164个半纤维素特征波长点。SiPLS-GSA经SiPLS优选的纤维素和半纤维素谱区波长点数分别为388个和160个,再经GSA进一步优选后得到157个纤维素特征波长点和148个半纤维素特征波长点。BiPLS-GSA经BiPLS优选的纤维素和半纤维素谱区波长点数分别为358个和180个,再经GSA进一步优选后得到130个纤维素特征波长点和153个半纤维素特征波长点。结果表明,通过波长优选,不仅参与建模的波长点数量显著减少,而且回归模型的性能显著优于全谱建模。其中,采用Full-GSA优选的纤维素特征光谱回归性能最佳,采用SiPLS-GSA优选的半纤维素特征光谱回归性能最佳。回归模型验证集的平均相对误差(MRE)分别为1.752 4%和2.020 8%,较全谱建模分别降低了13.636 6%和25.368 4%。基于结合温度参数设计适应度函数的策略构建的GSA具有良好的全局搜索性能,适用于玉米秸秆纤维素和半纤维素含量NIRS特征波长优选。GSA以全谱每个波长点为染色体基因的编码方案适用于NIRS全谱的特征波长优选。GSA同样适用于SiPLS和BiPLS优选后谱区的特征波长优选,能够有效实现优选后谱区的波长点优选。     

Simulated Annealing Using Hybrid Monte Carlo

We propose a variant of the simulated annealing method for optimization in the multivariate analysis of differentiable functions. The method uses global actualizations via the hybrid Monte Carlo algorithm in their generalized version for the proposal of new configurations. We show how this choice can improve upon the performance of simulated annealing methods (mainly when the number of variables is large) by allowing a more effective searching scheme and a faster annealing schedule.     

用模拟退火算法设计光学有限脉冲响应滤波器

提出了利用记忆型模拟回火退火算法来进行基于晶体双折射的光学有限脉冲响应滤波器设计的新方法，用以实现任意光谱响应双通道输出。模拟退火算法模拟了固体缓慢退火从而获得最低内能的物理过程，是一种简单而有效的随机性全局优化算法。应用具有记忆功能的回火退火算法，在较少的晶体延迟片数目下，实现了较优的设计。为证明该设计方法的可行性和有效性，给出了宽波带绿色／品红(Green／Magenta，G／M)滤波器的设计实例以及实验结果。初步研究表明，所设计结果完成满足LCOS投影分合色的要求。     

Remote sensing of atmospheric duct parameters using simulated annealing

Simulated annealing is one of the robust optimization schemes. Simulated annealing mimics the annealing process of the slow cooling of a heated metal to reach a stable minimum energy state. In this paper, we adopt simulated annealing to study the problem of the remote sensing of atmospheric duct parameters for two different geometries of propagation measurement. One is from a single emitter to an array of radio receivers (vertical measurements), and the other is from the radar clutter returns (horizontal measurements). Basic principles of simulated annealing and its applications to refractivity estimation are introduced. The performance of this method is validated using numerical experiments and field measurements collected at the East China Sea. The retrieved results demonstrate the feasibility of simulated annealing for near real-time atmospheric refractivity estimation. For comparison, the retrievals of the genetic algorithm are also presented. The comparisons indicate that the convergence speed of simulated annealing is faster than that of the genetic algorithm, while the anti-noise ability of the genetic algorithm is better than that of simulated annealing.     

Solving the undirected feedback vertex set problem by local search

An undirected graph consists of a set of vertices and a set of undirected edges between vertices. Such a graph may contain an abundant number of cycles, in which case a feedback vertex set (FVS) is a set of vertices intersecting with each of these cycles. Constructing a FVS of cardinality approaching the global minimum value is an optimization problem in the nondeterministic polynomial-complete complexity class, and therefore it might be extremely difficult for some large graph instances. In this paper we develop a simulated annealing local search algorithm for the undirected FVS problem by adapting the heuristic procedure of Galinier et al. [P. Galinier, E. Lemamou, M.W. Bouzidi, J. Heuristics 19, 797 (2013)], which worked for the directed FVS problem. By defining an order for the vertices outside the FVS, we replace the global cycle constraints by a set of local vertex constraints on this order. Under these local constraints the cardinality of the focal FVS is then gradually reduced by the simulated annealing dynamical process. We test this heuristic algorithm on large instances of Erdös-Rényi random graph and regular random graph, and find that this algorithm is comparable in performance to the belief propagation-guided decimation algorithm.     

模拟退火算法在波束图设计中的应用

摘要本文提出了基于模拟退火算法的波束设计方法。模拟退火算法模拟了固体缓慢退火从而获得最低内能的物理过程，是一种简单而有效的随机性全局优化算法。为证明该设计方法的可行性和有效性，给出了15元平面阵的设计实例以及仿真结果。结果表明，使用该方法设计的波束与常规方法相比，获得了更低的旁瓣级。     

A computed tomography reconstruction algorithm based on multipurpose optimal criterion and simulated annealing theory

Although emission spectral tomography (EST) combines emission spectral measurement with optical computed tomography (OCT), it is difficult to gain transient emission data from a large number of views,therefore, high precision OCT algorithms with few views ought to be studied for EST application. To improve the reconstruction precision in the case of few views, a new computed tomography reconstruction algorithm based on multipurpose optimal criterion and simulated annealing theory (multi-criterion simulated annealing reconstruction technique, MCSART) is proposed. This algorithm can suffice criterion of least squares, criterion of most uniformity, and criterion of most smoothness synchronously. We can get global optimal solution by MCSART algorithm with simulated annealing theory. The simulating experiment result shows that this algorithm is superior to the traditional algorithms under various noises.     

An emergent computation approach to the problem of polygon layout with performance constraints

We address the two-dimensional satellite irregular polygons layout problem (TDSIPLP). By mapping the layout problem to a physical system, we present a self-organization emergence algorithm (SEA) based on a multi-agent system (MAS). SEA uses information on the local energy of agents and neighborhood rules to guide interactions between the agents and the environment. System self-organization evolves at a microcosmic level and eventually yields a layout that meets the performance constraints for a satellite. Experiments show that SEA can provide much better performance than other optimization techniques, such as generalized simulated annealing (GSA).     

基于组合混沌遗传算法的认知无线电资源分配

提出了基于组合混沌遗传算法用于认知无线电资源分配,设计了相应的组合混沌序列发生器,并分别运用组合混沌遗传算法、粒子群优化算法、模拟退火算法及简单遗传算法对认知无线电资源分配问题进行了仿真分析.结果表明,组合混沌遗传算法具有收敛速度快、搜索空间广、全局收敛等优点.相比其他三种算法,基于组合混沌遗传算法进行资源分配提高了认知无线电系统的传输速率,降低了系统的发射功率及误码率,同时加快了收敛速度. 关键词： 组合混沌 认知无线电 遗传算法 资源分配     

基于模拟退火算法的宽角度X射线超反射镜设计研究

应用于硬X射线波段的宽带多层膜光学元件———宽角度X射线超反射镜的设计可以归结为一个连续变量的多维多极值的全局优化问题。缺少一种有效的全局优化方法是阻碍解决这一难题的一个关键。模拟退火算法是一种简单而且通用的全局优化算法。结合光学多层膜的设计原理提出了利用模拟退火算法来进行宽角度X射线超反射镜设计的新方法。结合已有的方法选择了W和C作为膜层的膜对材料,设计出Cu的Kα线处角度范围0.9°~1.1°反射率达到20%的宽角度X射线超反射镜。并在此基础上采用改进的自适应模拟退火算法实现了Cu的Kα线处宽角度X射线超反射镜的理想设计结果。设计结果表明了模拟退火算法在多层膜最优化设计领域的正确性和有效性。     

Diffusion model for annealing algorithm with perturbation quadratic approximation

We have dealt with the simulated annealing algorithm as a suitable method for finding the global minimum of a functionf(x) of one variable. The convergence properties of this algorithm have been investigated via diffusion equation solved at a fixed temperature. The convergence rate is mainly governed by the first subdominant eigenvalue of the generator. We have tried to calculate approximately this eigenvalue for arbitrary minimized functionf(x). These calculations lead us to propose some parameter for controlling the temperature.     

用改进的两步模拟退火法进行二元光学元件的设计

为了减少二元光学元件设计的计算量并提高设计精度,在对现有算法机理进行深入分析的基础上,提出了适用于二元光学元件设计的两步模拟退火法.该算法在整个退火过程中采用先量化后优化的策略,并将优化过程分为两个阶段：搜索并锁定最优解区间;快速收敛到最优解.模拟实验显示,与传统设计方法相比,该算法不仅保持了全局寻优的特点,而且提高了稳健性和效率.算法剔除了对设计结果影响较大的量化误差,提高了设计精度.用此法实例设计了单焦面辐射聚焦元件,得到了与目标图像一致的光学实验结果.     

Refractivity estimation from radar sea clutter

This paper addresses the problem of estimating lower atmospheric refractivity under the nonstandard propagation conditions frequently encountered in low altitude maritime radar applications. The vertical structure of the refractive environment is modeled by using a five-parameter model, and the horizontal structure is modeled as range-independent. The electromagnetic propagation in the troposphere is simulated by using a split-step fast Fourier transform based on parabolic approximation to the wave equation. A global search marked as a modified genetic algorithm (MGA) for the 5 environmental parameters is performed by using a genetic algorithm (GA) integrated with a simulated annealing technique. The retrieved results from simulated runs demonstrate the ability of this method to make atmospheric refractivity estimations. A comparison with the classical GA and the Bayesian Markov Chain Monte Carlo (Bayesian-MCMC) technique shows that the MGA can not only shorten the inverse time but also improve the inverse precision. For real data cases, the inversion values do not match the reference data very well. The inverted profile, however, can be used to synoptically describe the real refractive structure.     

基于混合算法的光学合成孔径成像系统孔径优化

结合模拟退火算法和遗传算法的优点,在遗传算法中加入模拟退火算子,运用这种混合算法对子孔径结构进行了优化。以阵列的子调制传递函数覆盖点间距最大化及最小冗余度来设计目标函数,实现了单圆周孔径阵列和双圆周孔径阵列的优化排列。使用仿真程序对优化阵列的光学传递函数进行了分析,将优化结果与相同子孔径数目的单圆周结构进行仿真实验和比较。结果表明,采用这种混合算法是可以很好的解决光学合成孔径结构的优化问题,优化结果接近全局最优解,优化后的双圆周阵列接近Golay阵列,具有更大的实际空间截止频率。     

基于遗传模拟退火算法的室内照明节能控制方法研究北大核心CSCD

目前的节能照明控制算法仍有陷入局部最优的问题。为了寻求全局最优解,提高室内照明的节能效果,设计一种遗传模拟退火算法对照明系统的控制参数进行优化求解。该算法通过在遗传操作后对优秀个体进行模拟退火处理,增强了算法的局部搜索能力。根据迭代的次数和种群的适应度对遗传概率进行自适应调节,使得算法在前期丰富种群多样性,避免算法“早熟”。提出基于人工神经网络的照度模型来计算室内照度分布,对照明舒适度进行评估,为构造优化算法的适应函数提供了依据。通过仿真实验,在本文介绍的照明场景应用遗传模拟退火算法,并与传统粒子群算法和遗传算法进行比较,其照明节能性能分别高出5.30%和13.61%。     

特形脉冲的设计与计算机模拟

根据LiouvillevonNeumann方程从理论上对特形脉冲做了全面的描述,提出了一种具体的调幅特形脉冲设计方案:首先将待设计的脉冲展成一个有限Fourier级数,然后根据Bloch方程的解析解准确计算出各阶正弦、余弦波的频谱,再将这些频谱组合后与该脉冲的理想频谱进行比较构成误差函数,最后运用鲍威尔-模拟退火组合优化算法计算出全局最优Fourier系数,即可得到所需脉冲的表达式.应用此设计方案,得到了体系处于热平衡态时的特形激励脉冲和反转脉冲的具体表达式.计算机模拟表明,所得脉冲的频谱具有较好的选择性 关键词： 核磁共振 特形脉冲 Bloch方程 鲍威尔模拟退火组合优化算法