关键词: 工作流, 访问控制, 资源分配, 约束满足

Abstract:

Exclusion constrained workflow satisfiability decision is an important problem that is concerned with the feasibility of secure business. Its representative algorithms so far are seriously out-of-balance between the theoretical and practical performance, or time and space cost. According to its low density of constraints, Jegou??s backtracking tree decomposition method is used to address the above issue. However, its concept of sub-problem independence is deduced only from the property of constraint irrelevance and has no guarantee to the compatibility between partial solutions. A complete sub-problem independence with a caching principle for partial solutions is set up from the view of both variable non-intersection and constraint irrelevance. A corresponding algorithm is designed and its correctness is proven via a method of alternating induction. Its time complexity is [O*(|S|3×dW+1)] which is conditionally lower than the most optimal time [O*(2|S|(|X|+|U|2))] so far, where S, d, W are the set of steps, the maximum scale of authorization lists for steps and the tree-decomposition width. Experiments show that, under the condition of low constraint density, the algorithm has remarkable time efficiency than the algorithms with the best theoretical or practical performance, while no quite much space cost is paid.

Key words: workflow, access control, resource allocation, constraint satisfaction