Optimal coordination of purchasing, inventory and demand management

採購, 庫存和需求管理的整合協調及優化模型

Student thesis: Master's Thesis

View graph of relations


  • Yi Kwan CHEN

Related Research Unit(s)


Awarding Institution
Award date15 Feb 2006


This research is motivated by the long-standing problems faced by one large wholesaler in Hong Kong, which plays the role as a sourcing agent for over a thousand of worldwide clients on hundreds of products. Around forty percent of the commodities are sold via contracts, and the remaining sixty percent are on an ex-stock basis, of which future demand patterns are very uncertain. The company has been facing several operational problems. Firstly, the company constantly suffers from short of warehouse space, incurring extra costs for acquiring public storage. Secondly, warehouse records the company has a very low inventory turnover rate, while in-house storage space is long occupied by low value and slow moving items, giving rise to cash being tied up and high opportunity cost. Thirdly, the product managers in the purchasing department find difficulties in determining the right order sizes based on historical data and forecasted demands, which often causes significant discrepancies between actual and planned sales, resulting in heavy shortage and inventory carrying costs as well as low overall efficiency. A further investigation into the system reveals the following fundamental causes of the above mentioned problems: (1) Each of the three main functional units, i.e. purchasing, warehousing and marketing, aims at its own operational goal and there is no shared objective for enhancing the overall system efficiency. (2) There is no cross-functional coordination throughout decision making processes. (3) The warehousing unit has no control over incoming and outgoing inventories of those slow moving products. (4) Within the purchasing department, products are grouped and managed separately by a number of product managers. However, there lacks coordination for product replenishments to minimize the overall system cost, especially when price discounts are available. (5) The lack of information sharing among functional units hinders communication and coordination of their decisions and activities. To resolve these problems, what the company needs is an integrated decision making mechanism that enables total system coordination and optimization. The core idea of this research is to develop such mechanism for effective coordinated decision making in purchasing, warehouse inventory and demand management. The overall framework is divided into two major stages: planning stage and control & implementation stage. In the planning stage, we model the strategic purchase scheduling under different scenarios, including: (1) base model, which covers the primary features of cross docking, multi-storage facilities and perishable products; (2) model with order aggregation, under which items from a common source or shipped by the same transportation mode can be jointly replenished to minimize the ordering and transportation costs. In view of discrepancies between planned demand and actual sales, which result in excess inventories or stock-outs, we therefore, propose another model named stock-pushing mechanism in the control & implementation stage to clear those excess inventories to improve overall inventory turnover. Given the system nature in this study, the model size and the solution time required will be dramatic. A two-phase hierarchical optimization model for solving realistic problems is developed and proved effective. This model consists of two phases: Meta level, which is used to determine the optimal resources allocation for different Meta groups. In the second phase, procurement schedule and quantities are determined for each product in the Product level within each Meta group. Numerical illustrations and extensive experimental testings are provided for validation and demonstration purposes. The significance of this research can be summarized into following major aspects. First, this study presents the first application of mixed-integer linear programming model in cross-functional integration and coordination for a wholesaler with simultaneous consideration of order aggregation, perishable goods, price discounts, cross-docking, multiple storage facilities and market clearing mechanism. Some studies for manufacturers with partial coverage of the above mentioned features can be found in the literature, but there is only a few seeking an overall optimal purchasing-warehousing-marketing model framework from a wholesaler’s perspective. Therefore, our study provides a good start for further studies on wholesaler in an integrated system. Second, as mentioned above, the integrated model proposed in this study allows consideration of full range operational features in a wholesaler’s concern. The concurrent handling of all these operational complexities in one framework, especially including the cross docking practice that has been receiving growing interests in the industry, not only contributes to this study, but also management improvement for other players who may involve similar functions. Third, we introduce an integrated framework to link demand management with the purchasing and warehousing planning and execution decisions, which has not been well studied in the existing literature. This stock-pushing mechanism provides a bridge to link up with the planning stage to tackle its insufficiencies and the uncertainties in the reality. Fourth, we examine an easy-to-implement two-phase hierarchical optimization model that provides a practical way to solve reasonably large size problems with such complexity on a regular PC. Fifth, this study provides a practical mean for the company to maximize its overall system performance. Numerical experiments in this study further prove the value and benefit of integration and coordination in operations management across the entire business process. The model and the methods introduced here can be easily extended to support other types of players, such as manufacturers, distributors and retailers with similar system settings.

    Research areas

  • Management, Business logistics