We study the classical multistage lot sizing problem that arises in distrib
ution and inventory systems. A celebrated result in this area is the 94% an
d 98% approximation guarantee provided by power-of-two policies. In this pa
per, we propose a simple randomized rounding algorithm to establish these p
erformance bounds. We use this new technique to extend several results for
the capacitated lot sizing problems to the case with submodular ordering co
st. For the joint replenishment problem under a fixed base period model, we
construct a 95.8% approximation algorithm to the (possibly dynamic) optima
l lot sizing policy. The policies constructed are stationary but not necess
arily of the power-of-two type. This shows that for the fixed based plannin
g model, the class of stationary policies is within 95.8% of the optimum, i
mproving on the previously best known 94% approximation guarantee.