Storage plays an important part in the field of logistics; however, conventional storage and operation may result in an overloaded condition due to low management efficiency. To address this problem, this study proposes the concept of temporary positions in interconnected warehouses. By analyzing the opportunity costs generated, a hierarchical optimization model was constructed to quantify the improvement of temporary positions. Firstly, this study analyses the characteristics of the warehouse operation and constructs a quadratic assignment model by considering safety and efficiency. Besides, a novel genetic-BP neural network was proposed to find the optimal route of a forklift. After that, a hierarchical optimization model was constructed by considering two types of opportunity costs. Finally, an empirical study demonstrates that the interconnected mode with temporary positions has a satisfactory performance for improving operational efficiency.