【作者】 张文涛； 双锴； 万能； 詹舒波； 苏森；

【Author】 ZHANG Wen-tao;SHUANG Kai;Wan Neng;ZHAN Shu-bo;SU Sen;State Key Laboratory of Networking and Switching,Beijing University of Posts &Telecommunications;Science and Technology on Information Transmission and Dissemination in Communication Networks Laboratory;Beijing University of Posts & Telecommunications;

【机构】 北京邮电大学网络与交换技术国家重点实验室； 通信网信息传输与分发技术重点实验室； 北京邮电大学；

【摘要】 多技能模型呼叫中心的座席资源被多个呼叫队列共享形成资源竞争,座席资源的互斥访问导致路由计算的并行程度低,因此互斥方法控制下的多进程系统对座席路由计算的性能难以提升,这成为云呼叫中心扩展能力的瓶颈。对此本文在对呼叫中心路由分配算法与资源竞争模型分析基础上,提出了一种并行的呼叫中心路由方法PCCRM,该方法通过临界资源粒度细化,将互斥范围从全局资源缩小到局部资源,可显著减少了进程间同步造成的延迟,提高了系统多节点并发性能。同时PCCRM引入有序竞争模型来保证该并行方法符合呼叫中心路由算法基本原则,满足功能性要求。实验结果表明PCCRM可以有效提升路由算法的多进程并行程度和系统整体性能。更多还原

【Abstract】 The agents in call center with multi-skilling model are shared by multiple queues which leads to resource competition and causes the routing algorithm cannot execute parallel for mutual exclusion issues.So it is hard to promote the performance of routing algorithm even with multiprocess based on mutual exclusion method which results in the poor scalability of Cloud Call Center.The article proposes a parallel method named PCCRM for Call Center based on the analysis of agent routing algorithm and resource competition.Through refinement of resource,PCCRM reduces the general mutual exclusion to local mutual exclusion.So PCCRM can reduce the process synchronization outstanding and promote the performance of multiprocess.Meantime an orderly competition model is introduced to guarantee the general principle of call center’s agents selecting algorithm.更多还原