Abstract: The reliability design problem of the automobile product subsystem in which the suppliers are involved is researched. The mathematic model in which the supplier creditability is considered includes two objectives: maximizing the subsystem reliability and the supplier creditability. The multi-objective model is transformed into the single objective nonlinear integer programming model by using a weighting method. This problem is solved by particle swarm optimization (PSO) algorithm where discrete particles are encoded to represent the “part-supplier” relationships. Then the self-adaptive objective function with a dynamic penalty factor is defined, which converts the optimization problem into the equivalent non-constraint problem. Such an objective function extends the search scope of particles to the near feasibility region and further improves the search speed and convergence performance as well. A case study on the reliability design optimization problem of the transmission system for the medium car is investigated. The original data about weights of the parts and the supplier creditability are obtained by applying the methods of quality function deployment and fuzzy synthesis evaluation. Simulation results demonstrate the practicability and the effectiveness of the proposed PSO.

Key words: Automobile product, Constraint handling, Particle swarm optimization algorithm, Reliability optimization, Supplier involved