摘要
计算机研究者大多采用不同的进化算法解决最小权三角剖分,但都存在收敛速度慢、且易于早熟的缺点。为此,通过分析原有蚁群模型的不足和蚁群算法求解最小权三角剖分存在的问题,提出采用改进的蚁群模型结合选择点集最大允许剖分的方法,为加快收敛和抵制早熟引入对角线调整机制形成新的融合算法。实验结果表明,该算法在收敛速度、收敛效果和计算时间上都优于现有算法。
To solve the defects of slow convergence and prematurity that is common in the known algorithms /'or Minimum Weight Triangulation(MWT), the conception and algorithm of the maximum of allowable triangulation for point set is proposed, thereby ant colony algorithm can be applied to MWT successfully. Diagonal exchange rule is put forward to speed up convergence ability of algorithm. Experimental result shows that the algorithm's the convergence velocity, the ability to resist prematurity to converge and run time is much better than the existing algorithms.
出处
《计算机工程》
CAS
CSCD
北大核心
2010年第22期197-199,共3页
Computer Engineering
基金
国家自然科学基金资助项目(60673102)
关键词
蚁群算法
最小权三角剖分
早熟
计算机视觉
信息素
Ant Colony Algorithm(ACA)
Minimum Weight Triangulation(MWT)
prematurity
computer vision
hormone