Optimal Reliability Allocation for Modular Software Systems Basis on Support Vector Clustering and Fuzzy Decision

Support vector clustering algorithm and fuzzy decision are applied to optimizing reliability allocation model for modular software systems in this paper, The factor set of reliability level and the weight of factor are established and the membership degree of the software module-the factor is obtained using expert information and SVM optimum clustering by analyzing the influential factors of reliability allocation, appointed reliability index of software system is distributed rationally to the module and further synthesis decision-making based on attachment degree of every factor. Both reliability of whole system are ensured and production cost and development period are reduced. The model is simulated and calculated by the example for software reliability of electric-hydraulic proportional control system, decision shows reliability of position control and real-time management module is the biggest, secondly parameter control module, initialization preprocessing module, fault-diagnosis module respectively.