PowerBank Rack & Cell Management

A 90-minute OOD simulation: implement a PowerBank with ordered capacity-constrained racks, time-dependent charged/depleted/spent cells, insertion, an equalization pass that swaps cells between adjacent racks, and a dispatch pass that drains charged cells while refilling vacancies. Requirements Implement a PowerBank containing ordered racks 1..k, from front to back. Each rack has its own capacity and holds cells identified by cellId. Each cell is loaded at a timestamp with a duration and moves through three time-derived states at dispatch time: charged: dispatchTimestamp < loadTimestamp + duration depleted: loadTimestamp + duration <= dispatchTimestamp < loadTimestamp + 2 * duration spent: loadTimestamp + 2 * duration < dispatchTimestamp; remove spent cells and treat them as nonexistent. The exact state at loadTimestamp + 2 * duration is not specified; clarify this boundary before coding. A cell's charge score is loadTimestamp + duration - dispatchTimestamp. Higher charge ranks first, with lexicographically smaller cellId as the tie-breaker. Implement three operations: Initialize the rack structure and per-rack capacities. loadCell(String cellId, double timestamp, double duration) -> boolean Insert the cell into the first rack, from front to back, with available capacity. Return true on insertion; return false if no rack can accept it. dispatchCell(double timestamp, int maxDispatchCells) -> List<String> Run an equalize phase, then a dispatch phase. Equalize from front to back. When a nonempty rack has a charged-cell ratio below 50% and the next rack contains a charged cell, swap the current rack's least-charged cell with the next rack's most-charged cell. Dispatch at most maxDispatchCells charged cells. Repeatedly take the most-charged cell from the earliest rack containing any charged cell. Fill a vacancy using the most-charged cell from a later rack, propagating refills from front to back one rack at a time. Emit each selected cell as {cellId}:{charged/depleted}. If fewer than maxDispatchCells charged cells are available, return all eligible results. Notes The prompt is reading-heavy: roughly 30 minutes can go into parsing the rules and another 10+ minutes into design. The difficulty is state management and selecting data structures that preserve the ordering and movement rules; there is no separate algorithmic trick, and the constraints are sufficient for a direct simulation. The boundary at exactly twice a cell's duration and the precise refill behavior across more than two racks should be confirmed before implementation. Preparation Translate the three time states, both tie-breakers, the 50% trigger, and the front-to-back movement rules into explicit invariants before writing code. Implement and test state transitions, per-rack ranked selection, equalization, and refill propagation independently before combining them into dispatchCell.