极大化提前完工总量平行机排序问题的LPT算法

研究带有共同交货期的三台平行机排序问题。工件在加工过程中不允许中断, 目标是极大化所有工件的提前完工量, 即在交货期前所加工工件(或部分) 的总加工时长。由于该问题是NP-难问题, 本文应用经典LPT算法来解决该问题。我们证明了LPT算法求解该问题的最坏情况界至多为$\frac{15}{13}$, 并给出实例说明最坏情况界的下界为$\frac{27}{25}$。     

Molecular Design in Practice: A Review of Selected Projects in a French Research Institute That Illustrates the Link between Chemical Biology and Medicinal Chemistry

Chemical biology and drug discovery are two scientific activities that pursue different goals but complement each other. The former is an interventional science that aims at understanding living systems through the modulation of its molecular components with compounds designed for this purpose. The latter is the art of designing drug candidates, i.e., molecules that act on selected molecular components of human beings and display, as a candidate treatment, the best reachable risk benefit ratio. In chemical biology, the compound is the means to understand biology, whereas in drug discovery, the compound is the goal. The toolbox they share includes biological and chemical analytic technologies, cell and whole-body imaging, and exploring the chemical space through state-of-the-art design and synthesis tools. In this article, we examine several tools shared by drug discovery and chemical biology through selected examples taken from research projects conducted in our institute in the last decade. These examples illustrate the design of chemical probes and tools to identify and validate new targets, to quantify target engagement in vitro and in vivo, to discover hits and to optimize pharmacokinetic properties with the control of compound concentration both spatially and temporally in the various biophases of a biological system.     

牛早期胚胎性别鉴定的PCR方法研究

探讨了ＰＣＲ法鉴定牛早期胚胎性别时所用的引物,反应缓冲液中Ｍｇ２＋浓度,循环反应的次数及可能造成污染的各种因素对性别鉴定结果的影响．并对１８个胚胎进行了性别鉴定,移植３８枚性别鉴定后的胚胎,成功产下２头与预测性别一致的牛犊,胚胎性别可鉴率为９４．１％,预测性别符合率为１００％．     

Constraints on the generalized natural inflation after Planck 2018

Based on the dynamics of single scalar field slow-roll inflation and the theory of reheating, we investigate the generalized natural inflationary (GNI) model. We introduce constraints on the scalar spectral index \begin{document}$n_{s}$\end{document} and the tensor-to scalar ratio r for the \begin{document}$\Lambda$\end{document}CDM \begin{document}$+r$\end{document} model, according to the latest data from Planck 2018 TT, TE, EE+low E+lensing (P18) and BICEP2/Keck 2015 season (BK15), i.e., with \begin{document}$n_{s}=0.9659\pm0.0044$\end{document} at 68% confidence level (CL), and \begin{document}$r<0.0623$\end{document} at 95% CL. We find that the GNI model is favored by P18 and BK15 in the ranges \begin{document}$\log_{10}(f/M_{p})= 0.62^{+0.17}_{-0.18}$\end{document} and \begin{document}$m=0.35^{+0.13}_{-0.23}$\end{document} at 68% CL. In addition, the corresponding predictions of generalized and two-phase reheating are discussed. It follows that the parameter m has significant effect on the model behavior.     

模糊模式识别在房地产预警信息系统中的应用

本文利用模糊模式识别和主成分分析等方法，进行房地产预警实证分析。以南京房地产市场为例说明。在面对若干繁冗驳杂的一系列指标时，如何进行指标归类，建立一套指标体系，以及如何分析这些指标反映的房地产市场所处的警情的层面。从而适时的得出一些分析结果，以供管理者做出及时的政策调整，避免房地产业危机的产生。     

Flow shop scheduling problems with deteriorating jobs on no-idle dominant machines

In this paper, we study the permutation flow shop scheduling problems with deteriorating jobs on no-idle dominant machines. The objective is to minimize one of the two regular performance criteria, namely, makespan and total completion time. For each objective, the following dominant machines constraint: idm, ddm, idm–ddm and ddm–idm are considered. We present a polynomial time solution algorithm for each of the four cases.     

K-Essential Phantom Energy: Revisited

We generalize some of those results which are closely related to the canonical kinetic term in k-essence formalism by further tuning the parameter (β). The scale factor a(t) could be negative and decreasing within a specific range of β (≡−1/ω, ω : the equation-of-state parameter) during the initial evolutional period.     

Rational selection of structurally diverse natural product scaffolds with favorable ADME properties for drug discovery

Natural product analogs are significant sources for therapeutic agents. To capitalize efficiently on the effective features of naturally occurring substances, a natural product-based library production platform has been devised at Aurigene for drug lead discovery. This approach combines the attractive biological and physicochemical properties of natural product scaffolds, provided by eons of natural selection, with the chemical diversity available from parallel synthetic methods. Virtual property analysis, using computational methods described here, guides the selection of a set of natural product scaffolds that are both structurally diverse and likely to have favorable pharmacokinetic properties. The experimental characterization of several in vitro ADME properties of twenty of these scaffolds, and of a small set of designed congeners based upon one scaffold, is also described. These data confirm that most of the scaffolds and the designed library members have properties favorable to their utilization for creating libraries of lead-like molecules.These authors have contributed equally to this work.     

Complexity results for scheduling chains on a single machine

We investigate the computational complexity of deterministic sequencing problems in which unit-time jobs have to be scheduled on a single machine subject to chain-like precedence constraints. NP-hardness is established for the cases in which the number of late jobs or the total weighted tardiness is to be minimized, and for several related problems involving the total weighted completion time criterion.     

Exact solution of a simple finite infinite source interaction model

Networks of queues with finite and infinite source customers have been used to study the interaction between the batch jobs and interactive jobs in computer systems. Earlier Kaufman ([1], PP-345-348) developed accurate approximations for a simple nonproduct form network of this type. In this paper we offer exact solutions for the same model with one finite source customer. We study both FIFO and LIFO disciplines at the contention node. The results are derived for the case where the finite source think time and service time distributions are generalized hyperexponential.