Production logistics(PL)is considered as a critical factor that affects the efficiency and cost of production operations in discrete manufacturing systems.To effectively utilize manufacturing big data to improve PL ef...Production logistics(PL)is considered as a critical factor that affects the efficiency and cost of production operations in discrete manufacturing systems.To effectively utilize manufacturing big data to improve PL efficiency and promote job shop floor economic benefits,this study proposes a PL trajectory analysis and optimization decision making method driven by a manufacturing task data chain(MTDC).First,the manufacturing task chain(MTC)is defined to characterize the discrete production process of a product.To handle manufacturing big data,the MTC data paradigm is designed,and the MTDC is established.Then,the logistics trajectory model is presented,where the various types of logistics trajectories are extracted using the MTC as the search engine for the MTDC.Based on this,a logistics efficiency evaluation indicator system is proposed to support the optimization decision making for the PL.Finally,a case study is applied to verify the proposed method,and the method determines the PL optimization decisions for PL efficiency without changing the layout and workshop equipment,which can assist managers in implementing the optimization decisions.展开更多
That the world is a global village is no longer news through the tremendous advancement in the Information Communication Technology (ICT). The metamorphosis of the human data storage and analysis from analogue through...That the world is a global village is no longer news through the tremendous advancement in the Information Communication Technology (ICT). The metamorphosis of the human data storage and analysis from analogue through the jaguars-loom mainframe computer to the present modern high power processing computers with sextillion bytes storage capacity has prompted discussion of Big Data concept as a tool in managing hitherto all human challenges of complex human system multiplier effects. The supply chain management (SCM) that deals with spatial service delivery that must be safe, efficient, reliable, cheap, transparent, and foreseeable to meet customers’ needs cannot but employ bid data tools in its operation. This study employs secondary data online to review the importance of big data in supply chain management and the levels of adoption in Nigeria. The study revealed that the application of big data tools in SCM and other industrial sectors is synonymous to human and national development. It is therefore recommended that both private and governmental bodies should key into e-transactions for easy data assemblage and analysis for profitable forecasting and policy formation.展开更多
联邦学习和群智学习等协作学习技术,能够在保护数据隐私的条件下充分利用分布在各地的分布式数据深度挖掘数据中所蕴含的知识,拥有非常广阔的应用前景,尤其是在强调隐私惯例和道德约束的医疗健康领域.任何协作工作都需要选择可靠的参与...联邦学习和群智学习等协作学习技术,能够在保护数据隐私的条件下充分利用分布在各地的分布式数据深度挖掘数据中所蕴含的知识,拥有非常广阔的应用前景,尤其是在强调隐私惯例和道德约束的医疗健康领域.任何协作工作都需要选择可靠的参与方,协作学习中全局模型的性能在很大程度上取决于参与方的选择.然而,现有研究在选择参与方时都没有对不同机构医疗数据中存在的异质性加以直接关注.导致包含稳定性在内的全局模型的性能难以得到保障.提出了从信誉的角度尝试探索求解该问题.以迭代协作学习的方式尽可能挑选出具有良好信誉的参与方进行协作学习,以获得稳定可靠的高性能全局模型.首先,提出了一个描述医疗机构数据质量的AI信誉值指标AMP(AI medical promise),以帮助其在医疗领域中形成良好的AI生态.其次,建立了一个基于后向选择的迭代协作学习(colback-learning)框架.在单次协作学习任务中,利用后向选择方法以多项式时间复杂度迭代计算出性能良好且稳定的全局模型,完成AMP计算和积累.在AMP信誉值计算中,制定了一个综合考虑全局性能指标的评分函数,以针对医疗领域更有效地指导全局模型的训练.最后,使用真实医疗数据模拟多样化的协作学习场景.实验表明,colback-learning能够选择可靠参与方训练得到性能良好的全局模型,模型的性能稳定性比现有最好的参与方选择方法提高了1.3~6倍.全局模型的可解释性与集中式学习保持了较高的一致性.展开更多
In this paper we consider sequences of observations that irregularly space at infrequent time in-tervals. We will discuss about one of the most important issues of stochastic processes, named Markov chains. We would r...In this paper we consider sequences of observations that irregularly space at infrequent time in-tervals. We will discuss about one of the most important issues of stochastic processes, named Markov chains. We would reconstruct the collected imperfect data as a Markov chain and obtain an algorithm for finding maximum likelihood estimate of transition matrix. This approach is known as EM algorithm, which includes main optimum advantages among other approaches, and consists of two phases: phase (maximization of target function). Continue the phase E and M to achieve the sequence convergence of matrix. Its limit is the optimal estimator. This algorithm, in contrast with other optimum algorithms which could be used for this purpose, is practicable in maximum likelihood estimate, and unlike to the methods which involve mathematical, is executable by computer. At the end we will survey the theoretical outcomes with numerical computation by using R software.展开更多
基金supported by The University Discipline(Professional)Top-notch Talent Academic Funding Project of Anhui Provincethe General Project of National Natural Science Foundation of Anhui Province.
文摘Production logistics(PL)is considered as a critical factor that affects the efficiency and cost of production operations in discrete manufacturing systems.To effectively utilize manufacturing big data to improve PL efficiency and promote job shop floor economic benefits,this study proposes a PL trajectory analysis and optimization decision making method driven by a manufacturing task data chain(MTDC).First,the manufacturing task chain(MTC)is defined to characterize the discrete production process of a product.To handle manufacturing big data,the MTC data paradigm is designed,and the MTDC is established.Then,the logistics trajectory model is presented,where the various types of logistics trajectories are extracted using the MTC as the search engine for the MTDC.Based on this,a logistics efficiency evaluation indicator system is proposed to support the optimization decision making for the PL.Finally,a case study is applied to verify the proposed method,and the method determines the PL optimization decisions for PL efficiency without changing the layout and workshop equipment,which can assist managers in implementing the optimization decisions.
文摘That the world is a global village is no longer news through the tremendous advancement in the Information Communication Technology (ICT). The metamorphosis of the human data storage and analysis from analogue through the jaguars-loom mainframe computer to the present modern high power processing computers with sextillion bytes storage capacity has prompted discussion of Big Data concept as a tool in managing hitherto all human challenges of complex human system multiplier effects. The supply chain management (SCM) that deals with spatial service delivery that must be safe, efficient, reliable, cheap, transparent, and foreseeable to meet customers’ needs cannot but employ bid data tools in its operation. This study employs secondary data online to review the importance of big data in supply chain management and the levels of adoption in Nigeria. The study revealed that the application of big data tools in SCM and other industrial sectors is synonymous to human and national development. It is therefore recommended that both private and governmental bodies should key into e-transactions for easy data assemblage and analysis for profitable forecasting and policy formation.
文摘联邦学习和群智学习等协作学习技术,能够在保护数据隐私的条件下充分利用分布在各地的分布式数据深度挖掘数据中所蕴含的知识,拥有非常广阔的应用前景,尤其是在强调隐私惯例和道德约束的医疗健康领域.任何协作工作都需要选择可靠的参与方,协作学习中全局模型的性能在很大程度上取决于参与方的选择.然而,现有研究在选择参与方时都没有对不同机构医疗数据中存在的异质性加以直接关注.导致包含稳定性在内的全局模型的性能难以得到保障.提出了从信誉的角度尝试探索求解该问题.以迭代协作学习的方式尽可能挑选出具有良好信誉的参与方进行协作学习,以获得稳定可靠的高性能全局模型.首先,提出了一个描述医疗机构数据质量的AI信誉值指标AMP(AI medical promise),以帮助其在医疗领域中形成良好的AI生态.其次,建立了一个基于后向选择的迭代协作学习(colback-learning)框架.在单次协作学习任务中,利用后向选择方法以多项式时间复杂度迭代计算出性能良好且稳定的全局模型,完成AMP计算和积累.在AMP信誉值计算中,制定了一个综合考虑全局性能指标的评分函数,以针对医疗领域更有效地指导全局模型的训练.最后,使用真实医疗数据模拟多样化的协作学习场景.实验表明,colback-learning能够选择可靠参与方训练得到性能良好的全局模型,模型的性能稳定性比现有最好的参与方选择方法提高了1.3~6倍.全局模型的可解释性与集中式学习保持了较高的一致性.
文摘In this paper we consider sequences of observations that irregularly space at infrequent time in-tervals. We will discuss about one of the most important issues of stochastic processes, named Markov chains. We would reconstruct the collected imperfect data as a Markov chain and obtain an algorithm for finding maximum likelihood estimate of transition matrix. This approach is known as EM algorithm, which includes main optimum advantages among other approaches, and consists of two phases: phase (maximization of target function). Continue the phase E and M to achieve the sequence convergence of matrix. Its limit is the optimal estimator. This algorithm, in contrast with other optimum algorithms which could be used for this purpose, is practicable in maximum likelihood estimate, and unlike to the methods which involve mathematical, is executable by computer. At the end we will survey the theoretical outcomes with numerical computation by using R software.
