Abstract: Furniture arrangement is an important part of virtual scene design, which can be widely applied to virtual reality, 3D games and interior design.Noticing that existing methods for placing furniture are prone to converge to local optimal solutions because of constraint confliction among cost terms, we present a hierarchical optimization strategy and adopt particle swarm optimization (PSO) to solve the problem.We first introduce a hierarchical tree to structurally organize the furniture in a scene so as to avoid constraint confliction as far as possible, and then employ PSO to realize the optimization.Considering the inherently parallelizable feature of PSO, we accelerate the PSO's solving based on GPU.Experimental results show that our method can improve the quality and efficiency of furniture arrangement.