基于整数线性规划的角色挖掘建模方法

常见的角色工程问题包括基本角色挖掘、边-角色挖掘、!-近似角色挖掘及最小噪声角色挖掘.四类问题的约束条件与优化目标存在异同点,而利用现有角色挖掘方法构建的角色工程系统过程烦琐,工程造价高昂.针对该问题,给出一种角色挖掘建模方法.该方法将四类角色挖掘问题分别转化为布尔矩阵分解问题,使用整数线性规划统一建模表示不同类型的角色挖掘问题,并给出角色挖掘算法描述.实例分析结果表明,该方法能够简化角色工程系统构建,降低挖掘成本.

Modeling Role Mining Using Integer Linear Programming

Basic role mining,edge role mining,!-approx role mining and min-noise role mining are conventional role mining problems. There are different constraints and objectives in these problems,besides constructing role engineering system by using the existing role mining methods is complex and costly. A method of modeling role mining was proposed,which transformed role mining problems into Boolean matrix decomposition problems. Integer linear programming was applied to model these problems,and the mining algorithm was proposed. Results of the application example showed that,the method not only simplifies construction process,but also reduces engineering cost.