For systems with multiple levels, such as communication systems, computing systems, control systems and critical power systems, redundancy allocation is usually considered in the design stage to improve the reliability of the system and ensure that the system can meet the performance requirements of customers. Whereas systems are subjected to failures and maintenance throughout their life cycle, system availability is considered more suitable for evaluating the performance of repairable systems. In this chapter, a general formula of system availability is derived and the simulation algorithm is also proposed to estimate the real system availability. During the design phase, not only the identical units can be considered redundant objects, but also a variety of alternative units with different specifications that can achieve the same function may be considered redundant objects. Therefore, this chapter aims to study an alternative non-homogeneous redundancy allocation problem in repairable hierarchical systems. Due to the large and complex solution space, along with the limitation of selecting only one level as the redundant object in the failure path, the search for an optimal solution becomes more challenging. Therefore, a particle swarm optimization algorithm combining a heuristic exploration operator is proposed to ensure the generation of feasible solutions in each generation. The numerical examples are carried out to evaluate the effectiveness of the simulation algorithm approach, and we also investigate the effect of model parameter changes on the optimal solutions through sensitivity analysis.