Abstract: To solve the fuzzy uncertainty problem of service resource evaluation in cloud manufacturing environment and optimize service recommendation effect,a cloud manufacturing service recommendation model based on multi-party heterogeneous evaluation of Quality of Service(QoS) and dual constraints of supply and demand was proposed.By considering the evaluation of multi-stakeholders such as service providers,cloud platforms and service users,and the heterogeneous evaluation forms such as numerical value,ratio,natural language and star,a multi-party heterogeneous evaluation system of six QoS indicators:service time (t),cost (c),availability (av),maintainability (ma),trustworthiness (tru) and satisfaction (sat) were constructed,which used emotion analysis method based on emotion dictionary to objectively measure heterogeneous evaluation.A dual constraint service recommendation model considered the matching requirements of the supplier and the demander was constructed to solve the multi-objective optimization problem and comprehensively evaluate the optimized service.A service recommendation list integrating objective optimization and subjective preference was obtained.The rationality and effectiveness of the proposed model were verified by a small remote control automobile processing service released on the cloud platform.

Key words: cloud manufacturing, quality of service, multi-stakeholder, heterogeneous, dual constraints of supply and demand, service recommendation

摘要： 为解决云制造环境下服务资源评价模糊不确定性问题,优化服务推荐效果,提出一个基于服务质量(Quality of Service,QoS)多方异质评价与供需双约束的云制造服务推荐模型。首先,考虑服务提供方、云平台与服务用户多利益相关方的服务评价,兼顾数值、比值、自然语言、打星等多种异质评价形式,并利用基于情感词典的情感分析等方法对异质评价进行客观衡量,构建了包含服务时间(t)、成本(c)、可用性(av)、可维护性(ma)、可信任性(tru)和满意度(sat)六大QoS指标的多方异质评价体系;其次,构建了考虑供需双方匹配要求的双约束服务推荐模型,对多目标优选问题进行求解并对优选服务进行综合评价,最终得到融合客观优选与主观偏好的服务推荐列表。最后,以在云平台发布的一款小型遥控汽车加工服务为例,验证了所提模型的合理性与有效性。

关键词: 云制造, 服务质量, 多利益相关方, 异质, 供需双约束, 服务推荐