a 0-1 linear programming method for optimal normal and pseudo parameter reductions of soft sets
a 0-1 linear programming method for optimal normal and pseudo parameter reductions of soft sets
复制标题
用于优化软集法向参数和伪参数约简的 0-1 线性规划方法
DOI：
10.1016/j.asoc.2016.08.052
复制
发表时间：
2017
影响因子：
8.7
通讯作者：
耿生玲
中科院分区：
文献类型：
--
作者：
韩邦合;李永明;耿生玲
本文旨在寻找求解软集参数约简问题的更好算法，并给出其潜在应用。首先，我们定义了主导支持度参数矩阵，并用它来解释任意一对对象选择值不同的本质原因。然后提出了将软集合的正规参数约简和伪参数约简问题转化为若干等价的0-1线性规划模型的技术，从而使软集合的约简问题可以用任何整数规划计算软件求解。与Ma等人提出的算法相比，实验结果表明，我们的方法是更有效的正常参数约简，特别是当参数的数量是大的。最后，我们建立了一个软件系统，展示了我们的方法在决策支持中的潜在应用。
This paper aims to find better algorithms for solving parameter reduction problems of soft sets and gives their potential applications. Firstly, we define the matrix of dominant support parameters and use it to explain the essential reasons for the different choice values of arbitrary pair of objects. Then we propose techniques for translating the normal and pseudo parameter reduction problems of soft sets into several equivalent 0–1 linear programming models, thus reduction problems of soft sets can be solved by any computational software for integer programming. Compared with the algorithm proposed by Ma et al., experimental results show that our method for normal parameter reduction is more efficient particularly when the number of parameters is big. At last we build a software system to show the potential applications of our methods in decision support.