供应链企业信任评估与可信服务组合方法研究

发布时间：2018-01-03 10:35

本文关键词：供应链企业信任评估与可信服务组合方法研究　出处：《燕山大学》2016年博士论文　论文类型：学位论文

【摘要】：随着科技水平的发展以及市场需求的不断变化,供应链中各企业间的合作越来越多,为了使企业成员都能受益,并且让每个企业都有较强的竞争力,就必须对供应链的构成及运作进行深入研究,在这样的大背景下,对供应链和供应链管理的研究逐渐成为热点。当前对企业合作伙伴关系问题的研究越来越受到重视,但研究主要是在强调一些技术和流程要素,而忽略了一个在长期合作中起着至关重要作用的因素—信任。在现有的研究中,对影响企业信任的因素、企业间信任的评估方法、企业可信服务的选择、可信服务组合等问题的研究还存在不足之处,针对以上问题,本文主要做了如下研究工作。首先,为提高企业交易信任评估的准确性,提出了一种企业交易综合信任评估方法。该方法从交易满意度、交易金额、交易次数、交易时间这四个影响交易信任的因素入手对直接交易信任进行了评估,并运用了奖惩机制来抑制企业的欺诈交易,而后引入了推荐信任机制,给出了基于推荐可靠度的推荐信任评估方法,并通过对两种信任值的融合来获取企业交易综合信任值,最后通过仿真实验验证了该方法的有效性。其次,为了提高企业信任评估的全面性和可行性,提出了一种全面评估企业信任的综合信任评估方法。该方法首先对企业本身基本属性所展现的信任进行了评估,构建了评价指标体系,以BP神经网络为基础给出了一种企业基础信任评估方法,实现了对企业基础信任的量化。结合企业交易综合信任和企业基础信任来获取企业综合信任值,并通过实例验证了该方法的可行性。再次,在对服务进行选择时,若存在虚假服务则会导致服务满意度的下降,为此提出了一种基于信任的服务选择方法。该方法以服务信任为基础对候选服务进行筛选,然后给出了一种基于服务质量、服务价格、服务时间及服务信任四个属性的模糊聚类方法。在此基础之上,通过设置服务属性的权重系数来实现客户对不同服务属性的偏好需求,根据优选服务策略选择综合服务评价最高的服务,可以使客户获得较高的服务满意度。仿真实验表明基于信任的个性偏好服务选择方法相对于基于信任的服务选择及随机服务选择在服务满意度方面有了较大的改善。最后,由于恶意服务参与服务组合会降低整个组合实际服务水平的问题,提出了一种基于信任的服务组合方法。该方对传统遗传算法进行了改进,在编码方面改进了原有的一维整数编码方案,并给出了一种自适应的交叉变异概率的设置方法,设计了动态适应度函数,并在此函数基础之上,结合服务信任给出了一种带自修正的动态适应度函数,通过修正可以使该服务的服务质量、服务价格及服务时间更接近于实际水平。仿真实验表明该方法在服务组合的时间效率和全局搜寻能力及实际适应度值相对于现有的方法有不同程度的提高。
[Abstract]:With the development of science and technology level and market demand, more and more cooperation between enterprises in the supply chain, in order to make the enterprise members can benefit, and let each enterprise has strong competitiveness, we must make up and operation of the supply chain for further research, in this context, the research on Supply Chain and supply chain management has gradually become a hot topic. The present study on enterprise partnership issues more and more attention, but the study is mainly focuses on some technical and process elements, while ignoring the one in the long term cooperation plays a vital role in the factors of trust. In the current study, the influence factors of enterprises trust, trust between enterprises, enterprise evaluation method, trustworthy service selection, there are still shortcomings in the research of trusted services composition problem, to solve the above problems, this paper mainly do the following research Research work. Firstly, in order to improve the accuracy of enterprise transaction trust evaluation methods, a comprehensive enterprise transaction trust evaluation is proposed. The method from the transaction satisfaction, transaction amount, transaction times, the transaction time of the four factors of the transaction trust direct transaction trust was evaluated, and the use of incentive mechanisms to curb fraud the transaction of enterprises, and then introduces the recommended trust mechanism, gives the recommendation trust evaluation method recommended based on reliability, and through the fusion of two kinds of trust value to obtain the comprehensive trust business transaction value, finally through the simulation experiments verify the effectiveness of the proposed method. Secondly, in order to improve the feasibility and comprehensive enterprise trust evaluation the method of comprehensive evaluation of enterprise comprehensive trust trust evaluation is proposed. Firstly, the enterprise itself shows the basic attributes of trust evaluation, construct The evaluation index system, based on BP neural network gives an enterprise based trust evaluation method, the quantitative enterprise foundation trust. To obtain the enterprise comprehensive trust value with business and enterprise comprehensive trust based trust, the feasibility of this method is verified by examples. Thirdly, in the selection of services. If there is false service would lead to reduced service satisfaction, this paper proposes a trust based service selection method. The method for screening of candidate services for service trust based, and provides a service based on service quality, price, service and Business Hours trust fuzzy clustering method. On the basis of the four properties above, the weight coefficient set to achieve customer service attributes of different service attributes preference, according to the preferred service strategy of integrated services of the highest evaluation The service, so that customers can get higher service satisfaction. Simulation results show that the individual preference service trust selection method with respect to service selection and random selection of trust services has a great improvement in the service satisfaction based on. Finally, due to malicious services in the service composition will reduce the actual level of service problems, put forward a service composition method based on trust. The side of the traditional genetic algorithm is improved in the aspect of encoding improved one dimensional integer encoding scheme of the original, and gives the method of setting an adaptive crossover and mutation probability, designed a dynamic fitness function based on function, combining with a given service trust self correcting dynamic fitness function, through the correction can make the service quality, service price and Business Hours more close to the real water Simulation results show that the time efficiency and global search ability and actual fitness of service composition in this method are improved to some extent compared with the existing methods.

【学位授予单位】：燕山大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：F274

【相似文献】