A two-phased semantic optimization modeling approach on supplier selection in eProcurement
Introduction
The proliferation of B2B e-Commerce in recent years has resulted in an explosion of eProcurement on the Internet. Procurement from various suppliers is a capital-intensive decision that often accounts for a large portion of the total operating costs (Bonser & Wu, 2001). Hence, it is very important to reduce purchase cost while selecting the right suppliers and it contributes to improve corporate competitiveness.
Research works related to supplier selection can be classified into two broad categories: a qualitative approach and a quantitative one. A majority of the research deals with qualitative supplier evaluation schemes. Given the economic importance and inherent complexity of the supplier selection process, only a few articles have addressed decision-making by quantitative methodologies. None of the supplier selection models, however, explicitly reflect the purchase policy or the supplier-related knowledge dynamically nor do most of them reflect the possibilities of purchasing several parts from a single supplier for price discount or bundling effect.
Purchase strategies depend on the situation of an organization. In order to support the strategies, diverse models are necessary. Recently, model warehouse (Bolloju, Khalifa & Turban, 2002) is one of the methods to solve these problems. However, it is not easy for purchaser to meet with diverse purchase strategies by using only several ready-made models, and it is very inefficient to prepare all combinations of models in advance from model management point of view. Moreover, candidate suppliers vary depending on purchase conditions. It is a little complex to build a goal model to select right suppliers among all of the potential suppliers. Hence, a simpler approach is needed to solve this problem.
In this research, we propose a two-phased semantic optimization modeling approach that formulates a goal model through model identification and candidate supplier screening for strategic supplier selection and allocation (SSA). In the procurement process, supply conditions of suppliers and purchase strategies of purchaser are considered together. Basically, a purchaser wants to minimize the purchase cost with supply conditions such as price discount and bundling while making purchase strategies. A purchase strategy affects a goal model. We build a goal model from SSA base models through the process of model modification that reflects purchase strategies.
Fig. 1 depicts the modeling architecture for SSA. The SSA procedure is broken down into goal model identification, candidate supplier screening, goal model formulation, and model solving. The details of each component are as follows.
- •
Goal Model Identification. A new specific goal model based on the SSA base model is identified by modeling factors, which compose a purchase strategy by purchase manager. Chang and Lee (2004) proposed three approaches to derive a goal model from a base model: the Primitive Model approach, the Full Model approach, and the Most Similar Model approach. The Primitive Model has only mandatory constraints and the Full Model has all possible requested constraints. The Most Similar Model is a model case that is the most similar one to a modeling request. Those models can be modified into a goal model by adding or deleting model components. Intuitively the Primitive Model Approach is effective when the goal model is similar to the Primitive Model (Chang & Lee, 2004). Thus, in this research, we use the Primitive Model Approach, because it starts from a simple model. The identified goal model is composed of an SSA Primitive Model and additional model constraints. The identified goal model can be represented as:
- •
Candidate Supplier Screening. Candidate suppliers are screened by the supplier screening factors, which compose a purchase strategy. The preliminary screened candidate suppliers must satisfy the purchaser's requirements for evaluation criteria such as quality, delivery, and price boundary. After screening, the goal model is described as:
- •
Goal Model Formulation and Model Solving. The Modeling components corresponding to a base model and added constraints in GMScreened formulate an optimization model GMOpt using model coefficients of candidate suppliers and it can be solved by an IP solver such as iLOG, LINGO, or LINDO, etc.
To describe the above approach, we organized this article as follows. In Section 2, we reviewed previous studies related to SSA. In Section 3, we introduced an SSA Primitive Model with price discount and bundling effect. In Section 4, we described two-phased model formulation in details. Finally, we conclude our study.
Section snippets
Related work
Purchasers in an organization buy many different types of items and services. The procedures used in completing a total transaction normally vary among the different types of purchases. Procurement is defined, in a narrow sense, as the act of buying goods and services for a firm or, from a broader perspective, as the activity of obtaining goods and services for the firm (Cavinato, 1984). Weele (1994) divided procurement process into five stages: identification of suppliers, supplier selection,
A Primitive Model for supplier selection and allocation
In this study, we introduce an SSA Primitive Model, which will be used as a base model for goal model formulation. We assume that suppliers supply their products with price discount (Chaudhry et al., 1993) and bundling (Rosenthal et al., 1995). Notations described in the SSA Primitive Model are as follows.
- i
index of item, i∈IOrd, where IOrd is a set of order items required by purchaser;
- j
index of supplier, j∈SCand[i], where SCand[i] is a set of candidate suppliers which sell item i;
- k
index of price
Goal model identification
Purchase strategies by purchaser managers are very diverse. Hence, each goal model generated from the SSA Primitive Model needs additional model constraints to cope with them. We considered them in three dimensions as depicted in Table 2: purchase quantity, supplier evaluation, and supplier selection. In the purchase quantity dimension, there are minimum and maximum order quantities. These mean that a purchaser sets constraints on quantity when he/she contracts with each supplier. The quality
Conclusions
The emerging digital economy is creating abundant opportunities for operations research (OR) applications such as financial services, electronic markets, network infrastructure, packaged OR software tools, supply chain management (SCM), and travel-related services (Geoffrion & Krishnan, 2001). In this paper, we applied OR to strategic SSA in B2B eProcurement and proposed an approach on SSA using the two-phased optimization model formulation. This approach formulates a goal model after semantic
References (20)
- et al.
Integrating knowledge management into enterprise environments for the next generation decision support
Decision Support Systems
(2002) - et al.
Case-based modification for optimization agents: AGENT-OPT
Decision Support Systems
(2004) - et al.
Vendor selection with price breaks
European Journal of Operational Research
(1993) - et al.
A multi-objective model for purchasing of bulk raw materials of a large-scale integrated steel plant
International Journal of Production Economics
(2003) - et al.
A decision support system for support selection using an integrated analytic hierarchy process and linear programming
International Journal of Production Economics
(1998) - et al.
The total cost of logistics in supplier selection, under conditions of multiple sourcing, multiple criteria and capacity constraint
International Journal of Production Economics
(2001) - et al.
A multiobjective approach to vendor selection
European Journal of Operational Research
(1993) - et al.
Logical representation of integer programming models
Decision Support Systems
(1996) - et al.
Improving purchasing productivity at IBM with a normative decision support system
Interfaces
(1985) - et al.
Procurement planning to maintain both short-term adaptiveness and long-term perspective
Management Science
(2001)