求解子旅行商问题的改进蚁群算法

已有求解子旅行商问题的蚁群算法存在容易早熟、易于陷入局部最优的问题。为此,提出一种改进的蚁群算法。将拥挤因子嵌入到蚁群算法的状态转移和信息素更新过程中,增强全局搜索能力,设计邻域搜索技术和局部变异技术,以提高解的质量和加快收敛速度。实验结果表明,该算法的求解质量和稳定性较好。     

基于心理物理学的弱视辅助检查方法

弱视是一种由于视觉系统发育障碍所引起,且不能通过光学途径以及手术加以矫正的视功能缺损,我国弱视检出率高达2.8%。弱视是一种发育性疾病,早期发现尽早治疗效果较好,因此,针对弱视的发病机制,借助心理物理学方法,设计一种检查方法帮助医生早期诊断病情,从而制定有效的个性化治疗方案具有重要的意义。运用Matlab7.0与Psychophysics Toolbox2.5设计并实现了以Gabor为刺激源,用于检查弱视患者对比敏感度和拥挤效应的新方法。经临床试用,这种检查方法操作简单、方便,对于进一步检查和诊断弱视患者的视功能缺损具有重要的意义。     

游人密度及道路宽度对城市绿地游览体验的影响

为了深入理解游人密度及道路宽度对城市绿地游览体验的影响,首先通过计算机技术创建30张包含2种道路宽度、5种游人密度、3种城市绿地类型的情景模拟图,再通过网络问卷的形式调查人们对这些场景的游览意愿。所得主要结论包括：1）在本研究中,无论何种城市绿地都需要一个适度的游人密度：当所见视域内人数为2人时,人工绿地拥有最佳的游览体验,而自然绿地的最佳游人人数为4人,高于人工绿地;2）道路拓宽对于人们的游览意愿的影响较小;3）总体来说,影响游览体验最主要因素是游人密度,当游客人数超出或低出最适度的游人密度后,游人的游览体验也随之降低。本研究为优化城市绿地游览体验的设计管理提供参考。     

环境规制对建筑业全要素生产率增长的影响研究

在利用Malmquist指数测算各地区建筑业全要素生产率(TFP)增长的基础上,使用动态面板系统GMM方法实证研究环境规制(ER)对建筑业TFP增长的影响。研究表明,ER对建筑业TFP增长存在"倒U"型的非线性影响:当ER强度较低时,其产生的"创新补偿效应"大于"挤出效应",推动TFP提升;当ER强度增大并超过临界点后,"挤出效应"大于"创新补偿效应",导致TFP下降。政府应合理控制环境规制强度,实现建筑业节能减排与TFP增长双赢的目标。     

适应值共享拥挤遗传算法

保持遗传算法在演化过程中的种群多样性,是将遗传算法成功应用于解决多峰优化问题和多目标优化问题的关键。适应值共享遗传算法和拥护遗传算法分别从不同角度改善了遗传算法的搜索能力,是寻找多个最优解的常用算法。将这两种算法的优点加以结合,提出适应值共享拥护遗传算法。数值测试结果表明,该算法比标准适应值共享遗传算法和确定性拥挤遗传算法具有更强的搜索能力。     

A multi-objective evolutionary algorithm based QoS routing in wireless mesh networks

The huge demand for real time services in wireless mesh networks (WMN) creates many challenging issues for providing quality of service (QoS). Designing of QoS routing protocols, which optimize the multiple objectives is computationally intractable. This paper proposes a new model for routing in WMN by using Modified Non-dominated Sorting Genetic Algorithm-II (MNSGA-II). The objectives which are considered here are the minimization of expected transmission count and the transmission delay. In order to retain the diversity in the non-dominated solutions, dynamic crowding distance (DCD) procedure is implemented in NSGA-II. The simulation is carried out in Network Simulator 2 (NS-2) and comparison is made using the metrics, expected transmission count and transmission delay by varying node mobility and by increasing number of nodes. It is observed that MNSGA-II improves the throughput and minimizes the transmission delay for varying number of nodes and higher mobility scenarios. The simulation clearly shows that MNSGA-II algorithm is certainly more suitable for solving multiobjective routing problem. A decision-making procedure based on analytic hierarchy process (AHP) has been adopted to find the best compromise solution from the set of Pareto-solutions obtained through MNSGA-II. The performance of MNSGA-II is compared with reference point based NSGA-II (R-NSGA-II) in terms of spread.     

基于新型拥挤度距离的多目标麻雀搜索算法

现实中的多目标问题日益复杂,解决这类问题需要高效的优化算法。基于麻雀搜索算法,提出多目标麻雀搜索算法（Multi-objective Sparrow Search Algorithm,MSSA）,对多目标优化问题进行求解。依据外部存档收敛性动态调整麻雀种群比例因子,以达到全局探索能力和局部开发能力的最佳平衡,确保收敛性;对麻雀种群进行非支配排序;对麻雀种群的发现者引入多项式变异因子,增强算法跳出局部最优的能力;设计一种新型拥挤度距离计算策略,利用外部存档解的拥挤度大小剔除相似个体的方法对种群进行裁剪,使个体不超过存档上限的同时维持种群的多样性。分别使用多目标函数和盘式制动器设计测试算法性能。MSSA与MOPSO、MOGWO、NSGA-II和SPEA2在多目标测试函数上进行对比实验,结果表明MSSA算法在收敛性和均匀性两项指标上有显著的优势。盘式制动器仿真结果表明,MSSA可以快速地找到问题的非支配解,证明了该方法的有效性。     

Thermal analysis of microwave GaN‐HEMTs in conventional and flip‐chip assemblies

Temperature rising which originates from self‐heating degrades the electrical characteristics, reliability, and lifetime of high‐power GaN‐HEMTs. In this article, a systematic analytical approach for thermal evaluation of microwave GaN‐HEMTs is constructed through combining and scrutinizing some of the basic static thermal analysis methods to provide a deeper insight into the process of the channel temperature rising and self‐heating with a much lower computational burden. The proposed systematic thermal analysis has been applied to two different assembly methods: conventional mounting and flip‐chip mounting. Although the mathematical and empirical equations used are simple enough to save time, the effects of several phenomena and different conditions on the channel temperature have been taken into account. These include heat crowding phenomenon, the effect of temperature‐dependent thermal conductivity of the transistor constituent materials, transistor geometry, die‐attach material properties, and bump dimensions. To validate the accuracy of the calculations, all the analytical analyses are followed by 3D thermal simulations in ANSYS software. The simulation results confirm the accuracy of the analytical calculations.     

基于前馈扰动的粒子群改进算法

使用线性系统理论分析了粒子群算法(Particle Swarm Optimization,PS0)陷入局部极值的原因。为使粒子种群跳出局部极值粒子重新获得活性,借鉴了鱼群算法中拥挤度因子的概念,提出了前馈扰动粒子群算法(Feedforward Disturbance Particle Swarm Optimization,FDPSO),在以当前最优值为圆心拥挤度因子为半径的圆域内统计粒子的数量,当粒子数量大于某一常数时候,认为种群将会陷入局部最优,因此提前给种群加入扰动。仿真实验证明了理论及所提出算法的有效性。     

Effects of trimethylamine N-oxide and urea on DNA duplex and G-quadruplex

We systematically investigated effects of molecular crowding with trimethylamine N-oxide (TMAO) as a zwitterionic and protective osmolyte and urea as a nonionic denaturing osmolyte on conformation and thermodynamics of the canonical DNA duplex and the non-canonical DNA G-quadruplex. It was found that TMAO and urea stabilized and destabilized, respectively, the G-quadruplex. On the other hand, these osmolytes generally destabilize the duplex; however, it was observed that osmolytes having the trimethylamine group stabilized the duplex at the lower concentrations because of a direct binding to a groove of the duplex. These results are useful not only to predict DNA structures and their thermodynamics under physiological environments in living cells, but also design of polymers and materials to regulate structure and stability of DNA sequences.