Sequential resource allocation with constraints: Two-customer case

We study a two-customer sequential resource allocation problem with equity constraint, which is reflected by a max–min objective. For finite discrete demand distribution, we give a sufficient and necessary condition under which the optimal solution has monotonicity property. However, this property never holds with unbounded discrete distribution.     

A possible human counterpart of the principle of increasing entropy

It is well-known that the principle of increasing entropy holds for isolated natural systems that contain non-adaptive molecules. Here we present, for the first time, an experimental evidence for a possible human counterpart of the principle in an isolated social system that involves adaptive humans. Our work shows that the human counterpart is valid even though interactions among humans in social systems are distinctly different from those among molecules in natural systems. Thus, it becomes possible to understand social systems from this natural principle, at least to some extent.     

Mobile relays for enhanced broadband connectivity in high speed train systems

With the introduction of wireless modems and smart phones, the passenger transport industry is witnessing a high demand to ensure not only the safety of the trains, but also to provide users with Internet access all the time inside the train. When the Mobile Terminal (MT) communicates directly with the Base Station (BS), it will experience a severe degradation in the Quality of Service due to the path loss and shadowing effects as the wireless signal is traveling through the train. In this paper, we study the performance in the case of relays placed on top of each train car. In the proposed approach, these relays communicate with the cellular BS on one hand, and with the MTs inside the train cars on the other hand, using the Long Term Evolution (LTE) cellular technology. A low complexity heuristic LTE radio resource management approach is proposed and compared to the Hungarian algorithm, both in the presence and absence of the relays. The presence of the relays is shown to lead to significant enhancements in the effective data rates of the MTs. In addition, the proposed resource management approach is shown to reach a performance close to the optimal Hungarian algorithm.     

一种基于STM32的NAND flash的块分配框架设计

STM32对大容量数据文件存储与管理问题可通过NAND flash来解决。而NAND flash的高效管理需要文件系统参与。NAND flash有特殊的块读写及擦除机制,一般的嵌入式文件系统组织结构并不完全兼容NAND flash。针对NAND flash的特点兼顾STM32的资源承受力要提出新的NAND flash块分配框架,框架通过块分配槽这种数据结构,在不使用块分配表与垃圾表的情况下,实现了NAND flash均衡负载与垃圾块的回收;同时通过节点分配栈与文件节点表的配合来提高STM32对文件的读写速度,空间利用率和系统性能。仿真实验和计算结果表明该块分配框架可有效提高NAND flash块的均衡负载与节省RAM空间。     

基于MetropolisQ学习的蜂窝网络在线信道分配

针对现有的蜂窝网络的在线动态分配模型具有的信道需求量大、呼叫动态变化时阻塞率高和收敛速度慢的缺点,设计了一种基于MetropoisQ学习的蜂窝网络的在线信道分配方法。首先,在考虑同信道限制、邻居信道限制和同小区限制的基础上,设计了在线信道分配的数学模型,然后在Q-Learning算法基础上的设计了一种基于资格迹的Q(λ)算法实现信道的在线分配,为了进一步提高收敛速度,采用Metropois规则对算法中动作的选择方式进行改进,实现探索和利用的平衡。为了验证文中方法,采用Matlab工具上进行实验,仿真实验结果表明文中方法能实现蜂窝通信网络的在线信道分配,且与其它方法比较,具有较少的信道需求量、较低的阻塞率和收敛速度,较其它方法具有较大优越性。     

Calcium Sulfite Solids Activated by Iron for Enhancing As(III) Oxidation in Water

Desulfurized gypsum (DG) as a soil modifier imparts it with bulk solid sulfite. The Fe(III)–sulfite process in the liquid phase has shown great potential for the rapid removal of As(III), but the performance and mechanism of this process using DG as a sulfite source in aqueous solution remains unclear. In this work, employing solid CaSO₃ as a source of SO₃²⁻, we have studied the effects of different conditions (e.g., pH, Fe dosage, sulfite dosage) on As(III) oxidation in the Fe(III)–CaSO₃ system. The results show that 72.1% of As(III) was removed from solution by centrifugal treatment for 60 min at near-neutral pH. Quenching experiments have indicated that oxidation efficiencies of As(III) are due at 67.5% to HO^•, 17.5% to SO₅^•− and 15% to SO₄^•−. This finding may have promising implications in developing a new cost-effective technology for the treatment of arsenic-containing water using DG.     

Copper Nanoparticles In鄄Situ Anchored on Nitrogen鄄Doped Carbon for High鄄Efficiency Oxygen Reduction Reaction Electrocatalyst北大核心CSCD

Compared with noble metal platinum (Pt)-based catalysts, inexpensive non-noble metal electrocatalysts have attracted extensive attention for oxygen reduction reaction (ORR). Herein, chitosan as a kind of biomass resource rich in nitrogen and carbon was used to prepare nitrogen-doped carbon (N-C) and N-C in-situ anchored by copper nanoparticles (Cu/N-C). The as-obtained N-C and Cu/N-C nanoparticles were successfully used as non-noble eletrocatalysts tested for ORR. Compared with the N-C, the Cu/N-C showed the high surface area of 607.3 m2窑g-1 with the mean pore size of 2.5 nm and the pore volume of 0.40 cm3窑g-1. The most positive Gibbs free energy change was the rate determining step for ORR process with the 4e mechanism, where the value of the Cu (111)/N-C(-0.39 eV) was lower than that of the N-C(-0.26 eV). The Cu/N-C exhibited superior onset and half-wave potentials (0.96 V and 0.84 V, respectively) in alkaline media (0.1 mol窑L-1 KOH), all of which are much better than those measured for N-C and commercial Pt/C. Furthermore, the Cu/N-C showed superior methanol crossover avoidance and oxygen reduction stability. © 2021 Authors. All rights reserved.     

高校实验室安全教育在线资源调查研究*

对国内高校实验室安全教育在线资源现状进行调研。调查结果表明:在线资源主要有慕课、虚拟仿真实验及安全教育考试系统。因此,在安全教育过程中,可充分利用现有线上资源,弥补线下教学因时间有限、容量有限导致的不足,从而提高安全教育实效。线上、线下安全教育如何有机结合从而最大限度地发挥其在高校安全教育中的作用,仍需不断探索。