基于无向图的专家群体决策会诊调度算法  

Expert Group Diagnosis Scheduling Algorithm Based on Undirected Graph

作　　者：徐儒[1,2] 何中市[1] 徐泽同[3] 

机构地区：[1]重庆大学计算机学院,重庆400030 [2]长江师范学院数学与计算机学院,重庆408100 [3]中国科学院数学研究所,北京100080

出　　处：《西南师范大学学报（自然科学版）》2011年第2期115-120,共6页Journal of Southwest China Normal University(Natural Science Edition)

基　　金：教育部"春晖计划"科研项目(Z2005-1-55003);重庆市科技攻关项目(CSTC;2008AB5038)

摘　　要：传统CAD的专家调度算法,大都集中在专家和患者一对一的会诊模式,或不同领域专家和患者多对一的会诊模式方面的研究,很少涉及对相同领域专家与患者多对一的会诊调度研究.针对这一问题,提出了群体决策思想的多专家会诊调度算法(MESDA),调度相同领域内的多位专家,为同一患者展开群体会诊.算法采用贪心策略,一方面考虑并行会诊的数量达到最大,另一方面考虑每一群诊项的专家成员数达到最大,通过滑动参数控制,从局部最优出发,逐步解决并行会诊和专家成员数的优化问题.实验结果表明:算法能够解决现实问题中相同领域专家和患者多对一的群体决策会诊的调度问题.Traditional CAD expert scheduling algorithm,is concentrated on one to one consultation of experts and patient model,or a different domain experts and the consultation patterns of patients with many to one study,very few experts in the field involving the same patients with many to one of consultation with Scheduling.Im view of this problem,it is proposed that the scheduling algorithm for multi-expert diagnosis（SAMED）,scheduling the same number of experts in the field,starting for the same patient group consultation.Algorithm uses the greedy strategy,on the one hand considering the maximum number of parallel consultations,on the other hand considering every member of a group of experts attending items to maximize the number,by sliding the parameter control,starting from a local optimum,and gradually solve the parallel consultation and expert membership optimization problem.Experimental results show that： the algorithm can solve practical problems,the same experts in the field,and consultation in patients with many to one of the group decision problem.

关 键 词：多专家会诊 群体决策 贪心策略 无向图 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]