On Improving the Dynamic Constraint Satisfaction Problems Repairing

Dynamic Constraint Satisfaction Problems (DCSPs) have been proposed to define and handle problems that change over time. To effectively repair DCSPs, Extended Partialorder Dynamic Backtracking (EPDB) uses Variable Ordering Heuristics (V OHs) and retroactive data structures, safety conditions and nogoods, which are saved during the search process. This paper shows that the drawback of EPDB is the exhaustive verification of orders, saved in safety conditions and nogoods, between variables. This verification affects remarkably search time, especially since orders are often indirectly deduced. Therefore, we propose an improved version of EPDB, the Improved Partial-order Dynamic Backtracking (IPDB), which is a fast version of EPDB insofar as it deduces orders directly, by supporting an additive structure of orders, that allows to have the information about all successors of each variable. We evaluate IPDB on various kinds of problems, with the use of different V OHs, and compare it, on the one hand, with EPDB to show its effectiveness compared to this approach, and on the other hand, with Maintaining Arc Consistency (MAC), the most widely used method and one of the most efficient generic approaches to solve Constraint Satisfaction Problems (CSPs), to conclude from what perturbation rate resolving a DCSP with an effective approach can be more advantageous than repair.