广义优先关系约束下Max-npv项目调度问题及其遗传算法

以往Max-npv项目调度问题的研究都假定活动之间的关系为单一结束-开始类型,现实中活动之间关系复杂多变,因此,将广义优先关系引入Max-npv项目调度问题中,构建了广义优先关系约束下的Max-npv项目调度模型。针对该优化模型设计了一种双层遗传算法,外层遗传算法负责任务执行模式的优化,内层遗传算法负责任务调度的优化。在内层遗传算法中,采用任务开始时间之差作为新的编码方式,大大简化了交叉变异算子,针对网络图中的环状结构设计了修复算子,确保了编码的有效性。通过一个算例对算法进行了测试,实验结果验证了算法的有效性。     

随机活动工期下求解资源约束项目最大期望净现值的动态调度算法

本文研究了随机活动工期下如何调度资源约束项目使得项目的期望净现值最大。首先对问题进行了界定,建立了相应的优化模型,其次针对问题的特点设计了一种动态规划算法。在算法设计的过程中,本文通过对项目网络图结构及不同状态最优值之间关系的分析,优化了动态规划算法状态的生成过程及状态最优值的求解过程,从而加快了算法的求解。使用随机生成的540个不同规模、不同结构的仿真案例对算法的有效性进行了验证,并分析了项目网络特征对算法效率的影响。实验发现：项目的次序强度对算法所需时间有着较大的影响,随着项目次序强度的减小,生成的状态数量会增加,从而计算时间也会增加。本文的研究可以为不确定环境下的项目调度提供决策支持。     

An ideal way of obtaining an optimal inspection permutation for a system with components connected in series

The problem of an inspection permutation or inspection strategy (first discussed in a research paper in 1989 and reviewed in another research paper in 1991) is revisited. The problem deals with an N‐component system whose times to failure are independent but not identically distributed random variables. Each of the failure times follows an exponential distribution. The components in the system are connected in series such that the failure of at least one component entails the failure of the system. Upon system failure, the components are inspected one after another in a hierarchical way (called an inspection permutation) until the component causing the system to fail is identified. The inspection of each component is a process that takes a non‐negligible amount of time and is performed at a cost. Once the faulty component is identified, it is repaired at a cost, and the repair process takes some time. After the repair, the system is good as new and is put back in operation. The inspection permutation that results in the maximum long run average net income per unit of time (for the undiscounted case) or maximum total discounted net income per unit of time (for the discounted case) is called the optimal inspection permutation/strategy. A way of determining an optimal inspection permutation in an easier fashion, taking advantage of the improvements in computer software, is proffered. Mathematica is used to showcase how the method works with the aid of a numerical example. Copyright © 2016 John Wiley & Sons, Ltd.     

???????????????????????????

论证了作图法的几何原理------四点共圆. 介绍了四点共圆作图法在计算机辅助下的求解 和构建复杂滑移线场的过程. 并以有心扇形场为例,演示了该方法的解题步骤. 通过实例计 算,将该方法的求解结果与作者用数值积分法和矩阵算子法的求解结果作了比较,证明该方 法计算精度高、绘图质量好.     

Extrapolation of isothermal retention data in gas chromatography for chiral recognition studies

Summary The calculation of capacity factors, k, from net retention times, t_R, and the corresponding dead times, t_M, at different temperatures suffers from the limited accuracy of the t_M values. If the temperature coefficient racy of the t_M values. If the temperature coefficient d ln k/d (1/T) only is required, it is sufficient to determine net retention times (t_R)_p at constant inlet pressure p_i for different temperatures, since the temperature dependence of (t_M)_p can be assumed as (t_M)_p=A·e^B/T, with B being approximately independent of the column inlet pressure and of the nature of the carrier gas. The extrapolation and interpolation of (t_R)_p may be either performed by linear regression or graphically with a nomogram for ln (t_R)_p versus 1/T. The resolution factor, , of two components, e.g. enantiomers which are resolved on a chiral stationary phase, can be treated in a similar way. Examples are given for the resolution of enantiomers of two non-proteinogenic amino acids on the new polysiloxane phase L-Chirasil-CPG.     

倾斜多孔腔体内瞬态自然对流的实验研究

本文对充满水和玻璃球的倾斜多孔介质腔体,在其一侧壁温维持恒定值、而相对侧壁温以此恒定值为时均值并按一定的幅值随时间周期性变化的情况下,实验研究了周期性壁温变化和腔体倾角对多孔介质腔体内的瞬态自然对流换热的影响.结果表明,通过多孔介质腔体的净热流总是从腔体的较低壁面指向较高壁面.     

带有资源柔性约束的max-NPV分布式多项目调度问题

现有的分布式资源约束多项目调度问题研究中,假定全局资源限量在多项目工期内不可突破且多以工期为优化目标。针对此问题,考虑全局资源可从外部获取,以净现值为目标,构建带有全局资源柔性约束的分布式多项目调度问题的整数规划模型并设计有效的求解算法。首先,界定问题并确定项目现金流的计算方法;然后,针对求解问题的NP-hard属性,设计了遗传-模拟退火混合算法(GA_SA)求解此模型。最后,通过多组数值实验,设计不同算法与GA_SA算法进行比较,并分析了关键参数对多项目净现值的影响。结果表明,GA_SA算法具有较好的求解效果;与传统的全局资源刚性约束条件相比,全局资源柔性使用状态可以显著改善分布式多项目的收益绩效。     

m阶马氏链关于广义随机选择系统的极限定理

主要研究广义随机选择系统中的m阶非齐次马氏链随机转移概率调和平均的a.s.收敛的强极限定理.在证明中采用了一种把网微分法与条件矩母函数相结合应用于马氏链强极限定理研究的新途径.作为推论,得到m阶非齐次马氏链随机条件概率一个公平比的强极限定理,并将已有的结果加以推广.     

两人进化博弈策略的决定机制与应用研究

通过建立一般情况下的两人进化博弈模型,给出了系统均衡点对应的矩阵行列式和迹表达式的经济含义,分析了16种典型情形下的进化稳定策略,详细讨论了均衡点稳定性分析结果所对应的博弈双方决策过程,从策略权衡的视角揭示了策略选择的内在机制。研究结果表明:不同策略前提下的相对净支付决定了系统的进化稳定策略,对方的策略选择以及自身可选策略的支付比较是影响博弈主体策略选择的两个基本要素,博弈主体会趋向于选择在对方策略既定下能够带来更大支付的策略。最后以环境治理中地方政府与企业以及地方政府之间的博弈关系为例,从对称博弈和非对称博弈两方面阐明了本文所构建模型在政策设计中的应用价值:针对不同案例,只要明确了两人博弈的支付矩阵,就可以通过计算相对净支付确定博弈双方的行为演化规律和稳定策略,从而简化计算过程,更加直接和更为便捷地为政策设计提供理论参考。     

When a compact (countably compact) set is closed. II

We obtain (a) necessary and sufficient conditions and (b) sufficient conditions for a compact (countably compact) set to be closed in products (sequential products) and subspaces (sequential subspaces) of normal spaces. As a consequence of these, sufficient conditions are obtained for (i) the closedness of arbitrary (countable) union of closed sets and (ii) the equality of the union of the closures and the closure of the union of arbitrary (countable) families of sets in these spaces. It is also shown that these results do not hold for quotients of even T ₄,-spaces.