We develop a category management model to aid retailers in the space c
onstrained decisions of which products to stock (assortment) and how m
uch shelf space to allocate to those products. The model is formulated
as a constrained optimization problem with two basic decision variabl
es: assortment and allocation of space to the items in the assortment.
The non-linearities in the objective function and the zero-one decisi
on variables disallow a closed form solution. We develop a heuristic s
olution procedure based on simulated annealing and test it on a proble
m with a known optimum. We also apply the technique to a larger proble
m without a known optimum. Finally, the solution found by simulated an
nealing is compared against a solution produced using a shelf allocati
on rule based on share of sales.