Research On Bid-sections Partition In Bidding Procurement Of Foreign Periodicals

Latest research achievements and developments around the world are usually published and communicated via foreign periodicals. Therefore, the procurement of foreign periodicals is an important part in the collection development of academic libraries. With the ever-increasing prices and types of foreign periodicals, libraries have to try every way to decrease average purchasing cost along with the commitment of purchasing quality in the purpose of getting more foreign periodicals with the same amount of money. Consequently, bidding procurement becomes the first choice for foreign periodicals purchasing as it has an advantage of attracting more suppliers. Although the practice has already realized the importance of partitioning bid-sections in bidding procurement, there are few reasonable and scientific approaches to partition foreign periodicals. Thus, this paper aims to investigate this problem and offer an efficient solution for partitioning bid-sections in the bidding procurement of foreign periodicals.Firstly, this paper analyzes the characteristics of foreign periodicals market, the potential advantages of partitioning bid-sections in bidding procurement and the current approaches adopted in literature resources procurement. Secondly, based on the theory studies, this paper focuses on clustering approach to partition foreign periodicals. After several interviews with experts in the field of foreign periodicals procurement, this paper extracts six indices to measure foreign periodicals from two perspectives: one is the potential agent suppliers'feedback of inquiry pricing and the other is the characteristics of foreign periodicals themselves. These indices include not only numeric ones but also categorical ones. Furthermore, the categorical indices lie in different hierarchies. As a result, there is no ready-to-use clustering program or software which can satisfy the special needs. Considering the above requirements, this paper proposes a new strategy to partition bid-sections of foreign periodicals based on K-prototypes clustering algorithm but with improvements on the initial prototypes selection and the categorical dissimilarity measure, then programs the improved algorithm with Java. Finally, this paper examines this strategy with the real case of foreign periodicals procurement in the library of Harbin Institute of Technology, 2008. The comparison analysis shows that the proposed approach can relatively make the bid-sections with similar sizes and periodicals in the same bid-sections with high similarity and purity on nation attribute. More importantly, it can decrease the bid-sections pricing index and probably lower the real cost. Thus, this strategy proves to be reasonable and deserves further practicing in real bidding procurement of foreign periodicals.