Constraint-directed search for all-interval series

摘要: All-interval series is a standard benchmark problem for constraint satisfaction search. An all-interval of size n permutation integers [0, n) such that the differences between adjacent are [1, n). Generating each an interesting challenge community. The very difficult in terms search space. Different approaches have been used to date generate all solutions AIS but space must be explored still remains huge. In this paper, we present constraint-directed backtracking-based tree algorithm performs efficient lazy checking rather than immediate propagation. Moreover, prove several key properties help prune significantly. reduced essentially results into fewer backtracking. We also scalable parallel versions our can exploit advantage having multi-core processors and even multiple computer systems. Our new generates up 27 while satisfiability-based state-of-the-art approach 24.