众包平台任务定价的一类新方法  

A New Method of Crowdsourcing Platform Task Pricing

作　　者：冯云乔 严灵毓[2] FENG Yun-qiao;YAN Ling-yu(Huadong China University of Petroleum,College of Science,Qingdao,266580;Hubei University of Technology,School of Computer Science,Wuhan,430068)

机构地区：[1]中国石油大学(华东)理学院,山东青岛266580 [2]湖北工业大学,计算机学院,湖北武汉430068

出　　处：《工业工程与管理》2018年第4期145-149,共5页Industrial Engineering and Management

基　　金：国家自然科学基金资助项目(61502155)

摘　　要：以众包平台中的“拍照赚钱”任务为例，分析影响这一任务定价和完成率的影响因素。采用更加精确的距离模型代替经纬度。通过建立logistic回归模型来拟合先前任务的完成率，建立了指数型多元回归定价模型。采用基于聚类分析的任务打包算法。将会员按信誉度高低排序，依次分发任务。借助双目标优化模型实现最优打包方案。仿真结果表明，该方法的任务完成增长率相对于传统方案增长了40．06％，总成本减少了12．549／6。Taking the ＂photographing and making money＂ task in the crowdsourcing platform as an example,we analysis factors that affect the pricing and completion rate of this task. A more accurate distance model is used to take place of latitude and longitude. Logistic regression and exponential multiple regression are used to get the concerning results of completion rate and the price. A packaging algorithm, based on clustering analysis method,is adopted in this step. Sorting the members according to their credibility, then distribute the tasks in turn. An optimal packaging resolution is generated by using a dual-objective optimization model. Simulation results show that task completion rate of the method has increased 40.06% and the total cost has reduced 12.54% compared to the traditional one.

关 键 词：任务定价 信誉度 聚类分析 双目标优化模型 

分 类 号：O21[理学—概率论与数理统计]