This paper develops optimal inventory models for a firm with multiple items and multiple branches by taking into account the price discount of sizable procurement and the scale economies of centralized warehousing and consolidated transportation. Four inventory strategies are proposed: independent ordering and decentralized warehousing, independent ordering and centralized warehousing, joint ordering and decentralized warehousing, and joint ordering and centralized warehousing. The performances of pure policy (all items subject to adopting only one of the four strategies) and mixed policy (different items allowed to adopting different strategies) are compared. The mixed policy employs genetic stepwise clustering and statistical agglomerative clustering methods to classify items into appropriate clusters and then to determine the best strategy for each cluster. A firm of four branches with fifty heterogeneous items is analyzed. It is found that, in terms of total cost minimization (including ordering, warehousing, transportation and procurement costs), the mixed policy with clustering is superior to pure policy without clustering. For mixed policy, genetic stepwise clustering performs better than statistical agglomerative clustering.
[1]
Frank T. Shu.
Economic Ordering Frequency for two Items Jointly Replenished
,
1971
.
[2]
R. Paul,et al.
Multi-Product Inventory Situations with One Restriction
,
1976
.
[3]
H. Pirkul,et al.
Capacitated Multiple Item Ordering Problem with Quantity Discounts
,
1985
.
[4]
Faruk Güder,et al.
Ordering policies for multi-item inventory systems subject to multiple resource constraints
,
1999,
Comput. Oper. Res..
[5]
Lin-Yu Tseng,et al.
A genetic approach to the automatic clustering problem
,
2001,
Pattern Recognit..
[6]
Lawrence W. Lan,et al.
Genetic clustering algorithms
,
2001,
Eur. J. Oper. Res..
[7]
S. Mondal,et al.
Multi-item fuzzy EOQ models using genetic algorithm
,
2003
.