精英蜂群算法及考虑利益相关者的众包定价模型  

Elitism bee colony algorithm and crowdsourcing pricing model based on stakeholders' preferences

作　　者：浦东平 樊重俊[1] 袁光辉 杨云鹏[1] Pu Dongping;Fan Chongjun;Yuan Guanghui;Yang Yunpeng(Business School, University of Shanghai for Science & Technology, Shanghai 200093, China;School of Information Management & Engineering, Shanghai University of Finance & Economics, Shanghai 200433, China;Laboratory Center, Shanghai University of Finance & Economics, Shanghai 200433, China)

机构地区：[1]上海理工大学管理学院,上海200093 [2]上海财经大学信息管理与工程学院,上海200433 [3]上海财经大学实验中心,上海200433

出　　处：《计算机应用研究》2019年第4期1039-1044,共6页Application Research of Computers

基　　金：国家自然科学基金资助项目(71774111);上海市教育委员会科研创新重点基(14ZZ131);上海市一流学科基金资助项目(S1205YLXK);沪江基金资助项目(A14006);沪江基金研究基地专项项目(D14008)

摘　　要：从众包活动参与者即主体企业、众包平台、平台会员的视角出发,研究服务众包定价问题。在考虑会员信誉度和任务聚集度的基础上,针对含有地理因素的众包任务设计打包分配定价方案。以服务成本、任务价值、会员收益等为导向,对不同任务进行组合配置,从而设计多目标规划任务配置及定价模型,并针对该模型构建了精英蜂群算法。在精英蜂群算法中,充分利用蜜源信息并着重考虑成长性较好的蜜蜂,进而避免了局部最优问题,提高了搜索效率。通过对众包服务企业运营数据进行分析,获取到众包服务会员特征及任务完成相关基础信息,以此进行仿真实验。仿真结果表明,通过众包任务打包定价机制,任务完成率、企业总成本、三方总收益等方面均有显著优化。综合模型及数据实验可知,众包任务在定价与发布过程中根据自身特征差异,在无差异服务中只需要考虑会员信誉度,对于具有服务差异性的任务则需要考虑打包发布。This paper studied the pricing problem of crowdsourcing model from the perspective of participants,which included enterprises,crowdsourcing platforms and members.On the basis of considering membership credibility and task aggregation degree,it designed an allocation pricing scheme for crowdsourcing tasks with geographical factor.Guided by service cost,task va-lue,member income,it constructed a multi objective programming model by combining different tasks,and designed the elitism bee colony algorithm for the model.In this algorithm,through the use of nectar source information and considering the improvement of bees,it significantly improved the search efficiency.Then analyzing E-commerce enterprise operation data,this paper got crowdsourcing membership characteristics,task completion rate and other basic information,and carried out the simulation experiment.Simulation results show that the task packaging pricing strategy has significant advantages in task completion rate,enterprise total cost,and the three parts total revenue.Through the model and data experiments,it can be seen that crowdsour-cing has its own characteristics in pricing and publishing.For nondifferentiated services,it only needs to consider the reputation of members,for differentiated services,the task assignment needs to be packaged.

关 键 词：服务众包 定价策略 任务发布 多目标规划 精英蜂群算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]