论文部分内容阅读
传统蚁群算法需利用显式评估函数引导算法搜索,因而不适用于那些优化性能指标不能或者难以数量化的系统.为此,提出一种将人对问题解的数量评价值作为目标函数值的交互式蚂蚁算法.从人机交互的特点出发,设计了算法模型的结构、信息素的放置方式与更新策略和用户的评价方式.最后利用模拟算法环境的函数优化实验和汽车造型草图设计实验进行了测试,测试结果表明所提出的算法具有较高的运行效率,并能较好地克服用户疲劳问题.
Traditional ant colony algorithm needs explicit evaluation function to guide algorithm search, so it is not suitable for those systems whose performance indexes can not be quantified or hard to be quantified.Therefore, a method is proposed to evaluate the number of human solutions to the problem as the objective function value Interactive ant algorithm.According to the characteristics of human-computer interaction, this paper designs the structure of the algorithm model, the pheromone placement method and update strategy and the user’s evaluation method.Finally, using the function optimization experiment of simulation algorithm environment and the car sketch design experiment The test, the test results show that the proposed algorithm has high operating efficiency, and can better overcome the user fatigue problem.