应用混合模型蚁群算法解决连续多自变量问题
|
摘要
蚁群优化算法自提出的十几年来已广泛应用于解决旅行商问题(tralvel salesman problem,简称TSP)等离散化的组合优化问题,随着对算法的进一步研究,近两年一些学者提出用该算法解决连续问题已经初见成效。进一步探讨了如何将该方法用于解决连续多自变量问题,通过对比文献[1]发现在当用该文献中提到的方法解决连续多自变量问题时,计算所得的解并非最优解,计算所得的目标函数值不准确。分析了出现问题的主要原因在于以往应用蚁群算法解决问题时大多采用基本蚁群算法的Ant-cycle system模型,在面对连续多自变量问题上是失效的。通过研究基本蚁群算法的原理,发现将Ant-cycle system模型与Ant-quantity system模型相结合,在解决连续多自变量问题时有很大突破。因而提出了一种新的蚁群算法模型——Ant-cycle&quantity system(ACQS模型),在此简称为混合模型算法。用该算法进行了大量的试验,取得了很好的结果。同时为了加快收敛速度,算法中还提出了信息素的奖惩机制,取得了很好的试验效果。
Ant colony algorithm has been used in assembled optimized questions such as TSP for more than ten years.With the deep research on this algorithm,some intestine researchers have fetched it in some simple continuous problems and got many optimistic results.The keystone of this paper is to resolve continuous and multi-variables questions.Comparing with the problem in reference and using the method provided by it we found that we couldn't get eximious results when facing the continuous and multi-variables questions.This paper believes that the main reason,which causes the former fault in continuous multi-variables,is that the "ant-cycle system" is mot suitable for the new problems.By studying the mechanism of basic ant colony algorithm,we put forward a new system of this algorithm——Ant-cycle﹠quantity system(ACQS);we call it as mixed system.Through a lot lf experiment with this system we got good results.We also bring forward the reward and punishment mechanism in the system in order to quicken the convergent speed,we obtain some excellent experimenting results.
Ant colony algorithm has been used in assembled optimized questions such as TSP for more than ten years.With the deep research on this algorithm,some intestine researchers have fetched it in some simple continuous problems and got many optimistic results.The keystone of this paper is to resolve continuous and multi-variables questions.Comparing with the problem in reference and using the method provided by it we found that we couldn't get eximious results when facing the continuous and multi-variables questions.This paper believes that the main reason,which causes the former fault in continuous multi-variables,is that the "ant-cycle system" is mot suitable for the new problems.By studying the mechanism of basic ant colony algorithm,we put forward a new system of this algorithm——Ant-cycle﹠quantity system(ACQS);we call it as mixed system.Through a lot lf experiment with this system we got good results.We also bring forward the reward and punishment mechanism in the system in order to quicken the convergent speed,we obtain some excellent experimenting results.
引文
[1]高尚,钟娟,莫述军.连续优化问题的蚁群算法研究[J].微机发展,2003,13(1):21-22.
[2]毕军,付梦印,张宇河.一种改进的蚁群算法求解最短路径问题[J].计算机工程与应用,2003(3):107-109.
[3]黎锁平,张秀媛,杨海波.人工蚁群算法理论及其在经典TSP问题中的实现[J].交通运输系统工程与信息,2002,2(2):54-57.
[4]熊伟清,余舜浩,魏平.用于求解函数优化的一个蚁群处算法设计[J].微电子学与计算机,2003(1):23-25.
[5]李艳君,吴铁军.求解混杂生产调度问题的嵌套混合蚁群算法[J].自动化学报,2003,29(1):95-101.
[6]周勇,陈洪亮.蚁群算法的研究现状及展望[J].微型电脑应用,2002,18(2):5-7.
[7]张纪会,高齐圣,徐心和.自适应蚁群算法[J].控制理论与应用,2000,17(1):1-3.
[8]陈双全,等.地震波阻抗反演的蚁群算法实现[J].石油物探,2005(6).
[2]毕军,付梦印,张宇河.一种改进的蚁群算法求解最短路径问题[J].计算机工程与应用,2003(3):107-109.
[3]黎锁平,张秀媛,杨海波.人工蚁群算法理论及其在经典TSP问题中的实现[J].交通运输系统工程与信息,2002,2(2):54-57.
[4]熊伟清,余舜浩,魏平.用于求解函数优化的一个蚁群处算法设计[J].微电子学与计算机,2003(1):23-25.
[5]李艳君,吴铁军.求解混杂生产调度问题的嵌套混合蚁群算法[J].自动化学报,2003,29(1):95-101.
[6]周勇,陈洪亮.蚁群算法的研究现状及展望[J].微型电脑应用,2002,18(2):5-7.
[7]张纪会,高齐圣,徐心和.自适应蚁群算法[J].控制理论与应用,2000,17(1):1-3.
[8]陈双全,等.地震波阻抗反演的蚁群算法实现[J].石油物探,2005(6).
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心