论文部分内容阅读
针对交互式遗传算法的核心问题——用户的疲劳问题,提出一种基于用户偏好模型的交互式遗传算法.在用户较偏爱个体的历史评价信息基础上,构建了基于部分的用户偏好模型,并据此结合考虑各构成部分权值及构成部分间关联关系对进化个体适应度值的影响,给出了一种更能符合用户情感偏好的适应值估算方法,以便在用户疲劳时由计算机有效地代替人工自动对新产生的进化个体进行评分.给出所提算法的关键技术与实现步骤,并将其扩展应用在汽车造型设计这一实践领域.与其他算法的实验结果对比验证了本文提出的算法具有较高运行性能,并能有效降低用户疲劳.
Aiming at the core problem of interactive genetic algorithm - the user’s fatigue problem, an interactive genetic algorithm based on user preference model is proposed.On the basis of user’s preference information of individual historical evaluation, a partial user preference model According to the above, we consider the impact of the weight of each component and the relationship between the components on the fitness value of evolutionary individuals, and give a method to estimate the fitness value which is more in line with the user’s emotional preference, so that when the user is tired, Instead of manually and automatically grading the newly generated evolutionary individuals.The key techniques and implementation steps of the proposed algorithm are given and extended to the practical field of automotive modeling design.Compared with the experimental results of other algorithms, The algorithm has high performance and can effectively reduce user fatigue.