Carousels are rotatable closed-loop storage systems for small items,
where items are stored in bins along the loop. An order at a carousel
consists of (say) n different items stored there. We analyze
two problems: (1) minimizing the total time to fill an order (travel
time) and (2) order delays as they arrive, are filled, and depart. We
define clumpy orders and the nearest-end-point heuristic
(NEPH) for picking them. We determine conditions for NEPH to be optimal
for problem (1), and under a weak stochastic assumption, we derive the
distribution of travel time. We compare NEPH with the nearest-item
heuristic. Under Poisson arrivals and assumptions much weaker than in
the literature, we show that problem (2) may be modeled as an
M/G/1 queue.