Software:
Links:
|
| A Simple Heuristic for Serial Inventory Systems with Fixed Order Cost |
|
|
| Abstract: |
We consider N-stage serial inventory systems with
fixed order costs at each stage. Customer demand can be deterministic or
stochastic. When demand is deterministic, an optimal policy must possess both
the nested and zero-inventory-ordering properties, i.e., the order quantities
satisfy an integer-ratio constraint. For stochastic systems, the form of the
optimal policies is unknown. Therefore, we focus on echelon (r,nq) policies
under which the base order quantities also satisfy an integer-ratio constraint.
In this paper, we develop a simple heuristic to calculate near-optimal control
parameters for both systems. The heuristic includes two steps. First, cluster
the stages based on ratios of cost parameters. Second, solve an EOQ problem or a
single-stage (r,nq) problem for each cluster. One feature for this heuristic is
that we do not convert the solution into a power-of-two policy. For the
deterministic system, we show the heuristic policy is 94%-effective. Numerical
studies show that this heuristic outperforms the power-of-two heuristic for
deterministic systems and those developed in Chen and Zheng (1998) for
stochastic systems. Our study indicates that managing the stochastic system is
very similar to managing the deterministic system. Thus, many insights from the
deterministic model can be carried over to its stochastic counterpart.
| |
| Keywords: |
simple heuristic, deterministic, stochastic,
fixed order costs | | |
| |
|
|
|
|
