A heuristic pathfinding algorithm for dynamic fault tolerance in manufacturing networks

Nowadays the increasing demand for high-reliability service compositions in manufacturing networks has brought new challenges for fault tolerance methods. It involves the real-time detection and rapid recovery of manufacturing services to deal with the unavoidable failures and errors. Appropriate dynamic fault tolerance methods need to be adopted to mask faults immediately after they occur in order to improve the reliability of the manufacturing network. Aiming at solving this problem, a dynamic fault tolerance method based on the pathfinding algorithm is thus put forward. First, a network model is constructed to explicitly describe the manufacturing services and their relationships. Then the dynamic fault tolerance problem can be modeled as a Multi-Constrained Optimal Path (MCOP) selection problem. On this basis, a novel Dynamic A* Search based Fault Tolerance (DAS_FT) algorithm is proposed to solve the NP-Complete MCOP problem. The propose d algorithm can find the suitable replacement schemes for failed service compositions with the help of the redundant resources in the manufacturing network, which will satisfy the Quality of Service (QoS) constraints of the manufacturing task at the same time. A set of computational experiments are designed to evaluate the proposed DAS_FT and other popular algorithms such as NSGA II and MFPB_HOSTP, which are applied to the same dataset. The results obtained illustrate that the DAS_FT algorithm can improve the reliability of the manufacturing network effectively. In addition, the DAS_FT can efficiently find the replacement schemes with better QoS compared with NSGA II and MFPB_HOSTP.