Answers are grounded in this paper and cite their sources.
The growth of e-commerce has increased demand for compact, responsive automated storage systems. Multi-level puzzle-based storage systems with autonomous shuttles offer high storage density, yet models that treat loading and unloading as instantaneous may underestimate retrieval times and misrepresent movement-handling interactions. This study develops a time-indexed optimization framework representing horizontal and vertical movement, explicit handling, collision avoidance, and completion at final unloading. A Time-Indexed Integrated Movement-Manipulation (TIMM) heuristic combines priority-based construction, local search, and independent schedule-replay validation. TIMM reproduced all seven proven optima and generated replay-valid schedules in 489 of 522 runs. All 60 runs for the tested low-density M200R20 configuration on the 10 × 10 × 10 grid were replay-valid. Across nine paired cases, one-period loading and unloading increased mean makespan from 5.56 to 7.67 (38.0%) relative to instantaneous handling. In five matched mobility cases, unrestricted vertical movement reduced mean makespan from 8.2 to 6.6 (19.5%). Shuttle scarcity increased makespan and reduced throughput. Closest-First outperformed Furthest-First under full shuttle availability, whereas neither rule dominated under scarcity. TIMM also matched the adapted ALNS-IP method in makespan on ten jointly solved instances while reducing median runtime from 13.836 to 0.415 s.
No takes yet. Share an insight, caveat, or question.
jneid et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: