多核共享资源冲突延迟上限优化方法

嵌入式多核结构的共享资源冲突是硬实时任务最差情况执行时间(worst-case execution time,WCET)估算的难点,而且通过减少共享资源冲突延迟的估算可以减少硬实时任务的WCET估算值,提高硬实时任务的可调度性。针对带有冲突感知总线(interference-aware bus arbiter,IABA)的嵌入式多核结构,提出了一种基于bank-column缓存划分的访存请求冲突延迟上限优化方法,根据bank冲突次数和冲突延迟上限的关系,该方法通过优化bank到核映射来减少bank冲突发生次数,从而减小冲突延迟上限和WCET估算值。实验结果表明,与现有冲突延迟上限界定方法相比,提出的方法能减少约29%的WCET估算值。     

抗干扰通信网端到端延迟上界仿真分析

考虑到通信网端到端的延迟抖动模式,为了降低数据传输耗时,进行抗干扰通信网端到端延迟上界仿真分析.根据MAC层中协议,计算抗干扰通信网端到端延迟,确定上界以及通信网端到端延迟抖动上界,在通信网式中模型中,引入漂移因子,完成通信网端到端延迟上界演算过程,实现抗干扰通信网端到端延迟上界仿真分析.仿真结果表明,上述方法的数据传输丢包率实验结果、通信网模型构建开销实验结果以及通信网模型构建与数据传输仿真耗时比重均较优,可为日后的通信网端到端服务提供参考.     

基于网络演算计算保证服务端到端延迟上界

归纳总结了网络演算,阐明了网络演算的两个基本工具——进入曲线和服务曲线,得出了服务曲线存在瓶颈效应、端到端延迟的理想与近似确定性上界、提供保证服务网络节点的服务曲线需求等结论,计算了服务曲线以速率等待时间及PGPS(packetizedgeneralizedprocessorsharing)形式表示的保证服务端到端延迟确定性上界.     

Improving the Evaluation of Generator Matrix G by Initial Upper Bound Estimation

STTC （space-time trellis code） can achieve both the diversity and coding gains. To maximize the advantages of STTC, two design criteria for slow Rayleigh fading channels will be used： i.e., the rank and determinant criteria. This paper focuses on the determinant criteria, which involves the evaluation of the generator matrix G. Evaluation is improved by pruning the search process earlier, which is made possible by estimating the initial upper bound prior to the search. In order to reduce the search complexity, the initial upper bound will be calculated at the minimal cycle. Comparatively, it can reduce the search space by 25.8%.     

利用时滞补偿器降低广义预测控制动态关联作用

本文提出了通过预加前、后纯时滞补偿器的方法降低多时滞多变量广义预测控制的闭环动态关联作用的思想,给出了设计方法,并讨论了设计参数的选择对系统的动态关联作用的影响,通过仿真证明了本文所提出的方法的有效性。     

一种快速收敛的最大置信上界探索方法

深度强化学习(Deep Reinforcement Learning,DRL)方法在大状态空间控制任务上取得了出色效果,探索问题一直是该领域的一个研究热点.现有探索算法存在盲目探索、学习慢等问题.针对以上问题,提出了一种快速收敛的最大置信上界探索(Upper Confidence Bound Exploration w...     

A New Upper Bound for the Cylinder Packing Problem

This paper introduces a new upper bound to the problem of fitting identical circles into a rectangle. This problem is usually referred to as the 'cylinder packing problem' or 'cylinder palletization'. In practice, it arises when it is desired to maximize the number of cylindrical items packed in an upright position onto a rectangle/pallet. The upper bound developed consists in determining the reduced pallet area by deducting a lower bound for the unused pallet area from the total area of the pallet. The upper bound for the number of identical circles to pack into the pallet is computed by the ratio reduced pallet area/circle area . The results obtained for five distinct sets of problems are analyzed and compared with previous bounds found in the published literature.     

分组延迟上限分析策略研究

分组延迟上限分析和控制是近年来研究的焦点之一。本文按照分析分组延迟上限所采用策略的不同,对延迟上限的分析策略进行了分类,分析、比较了各类分析策略的特点和优缺点。对未来的研究方向进行了展望,给出了若干值得研究的问题。     

An Upper Bound on the Loss from Approximate Optimal-Value Functions

Many reinforcement learning approaches can be formulated using the theory of Markov decision processes and the associated method of dynamic programming (DP). The value of this theoretical understanding, however, is tempered by many practical concerns. One important question is whether DP-based approaches that use function approximation rather than lookup tables can avoid catastrophic effects on performance. This note presents a result of Bertsekas (1987) which guarantees that small errors in the approximation of a task's optimal value function cannot produce arbitrarily bad performance when actions are selected by a greedy policy. We derive an upper bound on performance loss that is slightly tighter than that in Bertsekas (1987), and we show the extension of the bound to Q-learning (Watkins, 1989). These results provide a partial theoretical rationale for the approximation of value functions, an issue of great practical importance in reinforcement learning.     

Bent函数数目的一个最好可能上界

依据Bent函数的特征矩阵,利用组合数学中的容斥原理和图论中的子图概念,给出在流密码学中有重要应用的Bent函数数目的一个上界,计算实例说明该上界是最好可能的上界。