基于群体决策的多用户协同交互式遗传算法  

Interactive Genetic Algorithm Based on Collective Decision Making with Multi-user Collaboration

作　　者：郭广颂[1] 文振华[1] 郝国生[2] GUO Guangsong;WEN Zhenhua;HAO Guosheng(School of Mechatronics Engineering,Zhengzhou University of Aeronautics,Zhengzhou 450046,Chin;College of Computer Science and Technology,Jiangsu Normal University,Xuzhou 221116,China)

机构地区：[1]郑州航空工业管理学院机电工程学院,郑州450046 [2]江苏师范大学计算机科学与技术学院,徐州221116

出　　处：《电子与信息学报》2018年第9期2165-2172,共8页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(61673196);河南省科技攻关项目(172102210513);河南省高等学校重点科研项目(18A120012)~~

摘　　要：采用交互式遗传算法求解大数据信息检索问题时,为实现偏好信息的提取和优化,单用户需完成较多数量的人-机交互操作,由此易产生用户疲劳、算法搜索效率低的难题。对此,该文在算法中引入多用户并行策略,通过群体决策优势,提高样本利用效率。首先,根据优化目标性质确定共性化协同或个性化协同类型,基于用户浏览行为计算用户相似度和个体相似度。然后,通过共享偏好相似用户的偏好相似个体预测个体区间适应值。基于个体表现型相似度聚类,提出大规模种群个体"区间数-区间数"适应值赋值策略。最后,依据子代种群个体与父代种群最优个体的相似性,推荐用户最佳评价个体。将所提方法应用于装饰性墙壁纸选型问题,并与已有典型方法比较。结果表明,所提方法在推荐个体质量、减轻用户疲劳、提高搜索效率等方面均具有优越性。When using interactive genetic algorithm to solve big data information retrieval problem, single user needs to complete more human-machine interactive operation to achieve preference information extraction and optimization, thus it is easy to generate the problem of user fatigue and algorithm low efficiency. A multi-user strategy is introduced by making full use of the advantages of group decision to improve the sample utilization efficiency. First of all, multi-user collaborative type is devided into common collaboration or personalized collaboration according to the optimization goal which calculats user similarity and individual similarity based on user＇s browsing behaviors. Then, individuals＇ interval fitness is forecasted by sharing similar individual of similarity users. Based on phenotype similarity clustering, the large scale population individuals of ＂interval- interval＂ fitness assignment strategy is introduced. Finally, the best evaluation individual is recommended according to the similarities between offspring individuals and parent individuals. The proposed method is applied to decorative wallpaper design problem and is compared with existing typical methods. The experimental results confirm that the proposed algorithm has advantages in improving optimization quality and alleviating user fatigue while improving its efficiency in exploration.

关 键 词：遗传算法 交互 群体决策 多用户 协同 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]