【摘要】 传统算法在构造解的过程中,利用随机选择策略,这种选择策略使得进化速度较慢,正反馈原理旨在强化性能较好的解,却容易出现停滞现象。这是造成蚁群算法的不足之处的根本原因.因而我们从选择策略方面进行修改,我们采用确定性选择和随机选择相结合的选择策略,并且在搜索过程中动态地调整作确定性选择的概率当进化到一定代数后,进化方向已经基本确定,这时对路径上信息量作动态凋整。缩小最好和最差路径上的信息量的差距,并且适当加大随机选择的概率,以小于l对解空间的更完全搜索,从而可有效地克服基本蚁群算法的不足,此算法属于自适应算法。更多还原

【Abstract】 Traditional algorithm in the structure solution process, by the random selection strategy, the selection strategy of making a slow evolution, positive feedback priprinciple aimed at strengthening the performance of a better solution, but prone to stagnation. This is caused by the root causes of the shortcomings of ant colony algorithm. So we from a strategy to modify, we adopt deterministic selection and random selection combining selection strategy, and in the search process dynamically adjust for deterministic choice probability when the evolution to a certain algebra, evolutionary direction has been basically established, then on the amount of information on the path as dynamic full wither. Narrowing the gap between the best and worst path on the amount of information, and appropriate to increase the probability of randomly selected, to less than 1 of the solution space more complete search, which caneffectively overcome the deficiency of the basic ant colony algorithm, this algorithm is adaptive algorithm..更多还原