针对动态覆盖问题可以转化为多目标优化问题,提出一种解决多目标优化的连续空间蚁群算法(Continuous Space Ant Colony System,CSACS).该算法通过随机划分过程,对连续解空间划分为多个子空间,分别在不同子空间利用蚁群进行区域内以及区...针对动态覆盖问题可以转化为多目标优化问题,提出一种解决多目标优化的连续空间蚁群算法(Continuous Space Ant Colony System,CSACS).该算法通过随机划分过程,对连续解空间划分为多个子空间,分别在不同子空间利用蚁群进行区域内以及区域间搜索Pareto最优解,为了保证最优解的多样性,引入小生境策略进行Pareto最优解适应度更新.实验表明,在不同网络规模和迭代次数下,区域覆盖度和网络寿命相对于传统经典算法有较好改进.字数以250字以上为宜.请不要在摘要中引用参考文献和英文缩略语.展开更多
A new algorithm is presented by using the ant colony algorithm based on genetic method （ACG） to solve the continuous optimization problem. Each component has a seed set. The seed in the set has the value of componen...A new algorithm is presented by using the ant colony algorithm based on genetic method （ACG） to solve the continuous optimization problem. Each component has a seed set. The seed in the set has the value of component, trail information and fitness. The ant chooses a seed from the seed set with the possibility determined by trail information and fitness of the seed. The genetic method is used to form new solutions from the solutions got by the ants. Best solutions are selected to update the seeds in the sets and trail information of the seeds. In updating the trail information, a diffusion function is used to achieve the diffuseness of trail information. The new algorithm is tested with 8 different benchmark functions.展开更多
文摘针对动态覆盖问题可以转化为多目标优化问题,提出一种解决多目标优化的连续空间蚁群算法(Continuous Space Ant Colony System,CSACS).该算法通过随机划分过程,对连续解空间划分为多个子空间,分别在不同子空间利用蚁群进行区域内以及区域间搜索Pareto最优解,为了保证最优解的多样性,引入小生境策略进行Pareto最优解适应度更新.实验表明,在不同网络规模和迭代次数下,区域覆盖度和网络寿命相对于传统经典算法有较好改进.字数以250字以上为宜.请不要在摘要中引用参考文献和英文缩略语.
基金project supported by the National High-Technology Research and Development Program of China(Grant No.8632005AA642010)
文摘A new algorithm is presented by using the ant colony algorithm based on genetic method （ACG） to solve the continuous optimization problem. Each component has a seed set. The seed in the set has the value of component, trail information and fitness. The ant chooses a seed from the seed set with the possibility determined by trail information and fitness of the seed. The genetic method is used to form new solutions from the solutions got by the ants. Best solutions are selected to update the seeds in the sets and trail information of the seeds. In updating the trail information, a diffusion function is used to achieve the diffuseness of trail information. The new algorithm is tested with 8 different benchmark functions.
