Hybrid Learning with New Value Function for the Maximum Common Induced Subgraph Problem

Maximum Common Induced Subgraph (MCIS) is an important NP-hard problem with wide real-world applications. An efficient class of MCIS algorithms uses Branch-and-Bound (BnB), consisting in successively selecting vertices to match and pruning when it discovered that a solution better than the best found so far does not exist. The method essential for performance BnB. In this paper, we propose new value function hybrid selection strategy used reinforcement learning define vertex method, BnB algorithm, called McSplitDAL, MCIS. Extensive experiments show McSplitDAL significantly improves current algorithms, McSplit+LL McSplit+RL. empirical analysis also performed illustrate why are effective.