<p>There is a rising need for responsive and innovative compact storage solutions to handle fluctuating demand volumes. When compared to other traditional storage systems, Puzzle-based storage (PBS) systems are better suited to handle high demand in small or compact storage. However, achieving high throughput often remains a challenge for these systems. This paper studies a new paradigm of PBS proposed as a result of the collaboration between Logistics 4.0 lab and the Norwegian company Wheel.me. This solution, called Puzzle-Based Movable Rack (PBMR) system, enables achieving together a high storage density and high throughput, and it is characterized by storage racks mounted on autonomous wheels. These autonomous wheels can move in any direction including diagonal movements, which could result in high throughput. In this study, we employed a PBMR system with diagonal movement and a class-based storage policy. Under the above configurations, we identified optimal first zone percentages that can minimize the average travel time. Further, we optimized the system performance while understanding the impact of different input parameters such as capacity and shape of the system, and demand profile to the optimal first zone.</p>