改进交互式蚁群算法及其应用

Improved Interactive Ant Colony Algorithm and Its Application

交互式蚁群优化(interactive ant colony optimization,i ACO)是一种利用人来评价解的优劣而进行系统优化的技术,可以求解性能指标不能或者难以数量化的优化问题。分析了交互式蚁群优化模型面临的研究困难。针对Tanabe等人提出的交互式蚂蚁算法性能不足的问题,提出利用全局历史最优解进行信息素的更新,并将信息素限定在一定区间内的改进交互式蚁群优化算法,从人机交互角度讨论了解的构造方法和人的评价策略。最后,利用函数优化和汽车造型设计进行了实验,运行结果表明算法具有较高优化性能。

Interactive ant colony optimization (iACO) is a technique that optimizes target systems based on human evaluation. It can be used to solve the systems whose optimization indices are unable or difficult to be quantificated.Firstly, this paper analyzes the difficulties faced by iACO model. Nextly, aiming at the low performance of interactive ant system that is put forward by Tanabe et al., this paper proposes an improved iACO algorithm. The pheromone in the proposed model is updated with the best ant and limited to a certain range. The method of constructing solutions and evaluating strategies from the perspective of human-computer interaction are discussed. Finally, the results of the experiments of function optimization and car styling draft design show that the proposed algorithm has higher optimization performance.