结合全局与局部信息的主动轮廓分割模型

针对传统的基于区域的主动轮廓模型在分割灰度不均匀图像和噪声图像存在效果不佳的问题,提出结合全局项与局部项的主动轮廓分割模型。全局项由CV(Chan-Vese)模型的保真项构成,局部项的构建考虑局部区域信息的同时引入反映图像灰度特性的局部熵信息。依据图像灰度的特点,选择合理的全局项和局部项参数,并加入正则项保证曲线在演化过程中保持平滑,保障分割结果的可靠性。通过变分水平集方法最小化能量泛函,依据梯度下降流迭代更新水平集,完成曲线演化。采用模拟图像和实际图像进行实验分析,结果表明,所提出的结合全局项和局部项的主动轮廓模型可以高效地分割噪声严重以及灰度分布不均匀的图像。     

在线测量精密机床和加工工件直线度的新方法——优化误差分离法（OEST)

目前,运用多传感器对工件和精密机床的直线度进行在线测量的方法有ＳＴＰ法、ＳＴＲＰ法和ＦＳＴＰ法等。由于受到传感器相邻间隔距离的限制,这些方法往往只适合于长工件的在线测量。本文推荐的新方法──优化误差分离法（ＯＥＳＴ）可以用于短工件中的测量,它是基于频域分析和时域优化搜索的原理,可以克服频域变换中出现的边缘失效问题以保证频域变换的正确性。本文还介绍了计算机仿真和实际测量实验的结果,并验证了该方法的可行性和正确性。     

基于Monte Carlo法的地空导弹火力分配优化问题研究

在确定地空导弹对目标的可射击系数、火力重叠修正系数和转火矛盾修正系数的基础上,结合Monte Carlo法的基本原理,介绍了解决地空导弹火力分配优化问题的一般步骤。运用计算机模拟技术,通过算例,证明了该方法的可行性,为有效解决地空导弹火力分配优化问题提供了一种较为科学的方法。     

基于改进局部投影算法的非线性时间序列降噪

改进了非线性时间序列降噪的局部投影算法,并应用此算法对含噪Lorenz混沌信号进行降噪,将其与小波分析降噪效果进行比较,实验结果表明:改进的局部投影算法对非线性信号降噪效果十分明显。     

科学制定发展规划 彰显地方高校职能

地方高校制定发展规划是适应国内外高等教育发展的客观需求，是区域经济社会发展的必要需求，是高校自身发展的迫切需求。地方高校要履行高校职能，更好地为地方经济社会发展服务就必须科学制定规划，特别是要战略规划学校的办学定位、特色和目标。同时，规划的实施要强调监控并要与高校职能协调统一。     

Counterinsurgency Challenge in Post-2001 Afghanistan

Although the Taliban insurgency was internally divided and unable to coordinate its activities in 2014–2015, the Afghan security forces were not able to contain it and steadily lost ground throughout 2015. Until 2015, there had been little effort to develop an indigenous Afghan counterinsurgency strategy, but a sense of urgency emerged after a string of Taliban victories. At the beginning of 2016, it was still not clear if and when the National Unity Government would be able to produce a counterinsurgency strategy and, in any case, the need for a coherent counterinsurgency approach became questionable as the Taliban appeared to be transitioning towards conventional warfare.     

Too much Western bias? The need for a more culturally adaptable approach to post-conflict security sector reform

The existing guidelines for security sector reform (SSR) tend to draw on theoretical work in the field of civil–military relations, which in turn has been derived from Western, liberal democratic models of governance. Although guidelines strongly advise that local culture and context need to be considered when drawing up objectives for post-conflict SSR programmes, this is not often reflected in practice. This article considers some of the reasons for this, citing both in-country challenges and donor-related issues, and suggests that one of the biggest problems is a lack of alternative, non-orthodox models of civil–military relations to draw upon. It is further suggested that elements of suitable alternative models may be found in states which possess political structures not entirely dissimilar to the Western, liberal democratic ideal, but which can offer different perspectives. Detailed research of these structures should produce a pool of sub-models which could then be employed to create bespoke, culturally appropriate objectives for use in post-conflict SSR programmes.     

Security sector reform in Africa: Donor approaches versus local needs

Many African states have security sector reform (SSR) programs. These are often internationally funded. But how do such programs account for previously existing security institutions and the security needs of local communities? This article examines SSR all over Africa to assess local ownership and path dependency from a New Institutionalist perspective. It finds that SSR, particularly in post-conflict countries, tends to be driven by ideas and perceptions of international donors promoting generalized blueprints. Often, such programs only account in a very limited way for path-dependent aspects of security institutions or the local context. Hence, the reforms often lack local participation and are thus not accepted by the local community eventually.     

基于响应面与遗传算法的柴油机喷射参数优化

为解决燃油喷射系统参数优化中高效寻求全局最优解问题,综合响应面设计和遗传算法的优点和不足,采用多项式响应面模型代替原始遗传算法中计算量庞大的目标特性分析模型,建立了多项式响应面模型和遗传算法相结合的参数优化方法。结合非线性规划的凸性分析证明了遗传算法在预定优化区间内全局最优解的唯一性。对优化匹配后的喷射参数进行了模拟计算,结果表明:优化后柴油机油耗降低了1.96g/(kW·h),功率上升了0.99%,排气温度降低了21.86℃。表明该方法具有简单可靠、优化效率高等特点。     

求解CVRP问题的一种改进启发式蚁群算法

针对蚁群算法求解CVRP问题时收敛速度慢、求解质量不高的缺点,提出了一种改进启发式蚁群算法。该算法借鉴蚁群系统和基于排列的蚂蚁系统的优点设计信息素更新策略,既加强了对每次迭代最好解的利用,又避免了陷入局部最优;按一定比例使用基本方法和基于PFIH方法构造路径,扩大了算法的搜索空间;采用一种混合局部搜索算子,增强了算法局部寻优能力。实验结果表明,改进启发式蚁群算法可以大幅度减少车辆运行成本,具有较快的收敛速度。