改进的基因表达式编程在符号回归中的众包定价分析

Improved gene expression programming in symbolic regression of crowdsourcing pricing analysis

移动互联网成为至关重要的信息获取终端,运用众包模式可为企业提供渠道智能监测优化服务,合理的任务定价对众包平台的科学管理和高效运行举足轻重,而研究任务定价规律是改善其定价合理性的前提。针对研究任务定价规律,提出父子两代相互竞争的选择策略,并提出“殖民策略”来产生子代种群,从而建立改进的基因表达式编程算法。实验结果为改进后的算法种群40代收敛,原算法种群180代收敛,且改进后的算法更易跳出局部最优解,从而证明改进后的算法有更快的收敛速度和更高的稳定性。最后,在考虑会员信誉度和任务聚集度的基础上,结合改进的基因表达式编程算法和符号回归理论,得出了广州和深圳的任务定价规律,增强了定价的合理性。

As Internet becomes a crucial information acquisition terminal,the use of crowd-sourcing mode provides enterprises with intelligent monitoring and optimization services for channels so that reasonable task pricing plays an important role in the scientific management and efficient operation of the crowd-sourcing platform.Studying the pricing law of task is the first and foremost step.In order to study the rule of pricing law,the selection strategy of father and son competing with each other was put forward,and the“colonial strategy”was built to generate the offspring population,so as to establish the improved gene expression programming algorithm.The experimental results show that the improved algorithm converged after 40 generations,the original algorithm converged after 180 generations,and the improved algorithm was easier to jump out of the local optimal solution.It is proved that the improved algorithm has faster convergence speed and higher stability.Finally,by considering member credibility and task aggregation,and combining with improved gene expression programming algorithm and symbolic regression theory,the task pricing rules of Guangzhou and Shenzhen were obtained,which enhances the rationality of pricing and gives a pricing scheme for new projects.