This research attempts to develop a decision support system for order selection. The proposed system is able to integrate both the quantitative and qualitative factors together. For the qualitative factors, the fuzzy IF–THEN rules are summarized from the questionnaire survey for the production experts and learned by a proposed fuzzy neural network (FNN) with initial weights generated by real-coded genetic algorithm (GA). Then, a feedforward artificial neural network (ANN) with error back-propagation (EBP) learning algorithm is employed to integrate the above two parts together. Both the simulation and real-life problem provided by an internationally OEM company results show that the proposed FNN can well learn the fuzzy IF–THEN rules. In addition, real-coded GA is proved to be better than the binary GA both in speed and accuracy. Considering both the quantitative and qualitative factors has more accurate results compared with considering only the quantitative factors
The improvement of information technology and the introduction of Internet, have made the world community so close like the neighborhood. This causes that the business model became completely different from what we were familiar with. Most of the companies can reach any place around the world only through a ‘mouse click’. Such an environment makes EC possible, which aims to achieve the quick response for all the customers. For supply chain management (SCM), which is a part of electronic commerce (EC), in order to enhance the commercial competitive advantage in a constantly fluctuating environment for all the supply chain partners, the managers of an organization must make the right decision in time depending on the information from the upper or lower stream companies. However, the decision lead-time ranges from several years to several hours based on the types of business. Thus, making an accurate decision in time plays a prominent role in EC.
Among the decisions critical to the managers, available-to-promise is supposed to be the first as well as the most significant decision. Conventional approach depends on the experience of production managers and the decision is made manually. Thus, how to automatically accept, or reject, the orders coming from all over the world becomes a tough job for the production managers in SCM, since each order posses its own characteristics. Not all the orders have the same contributions to the company. In order to overcome this practical problem by providing the production managers a real-time solution, this research intends to summarize the experiences of production experts on order selection as the fuzzy IF–THEN rule base. Then, they are combined with the quantitative factors via an ANN.
Therefore, this research is dedicated to developing a decision support system for order selection. The proposed system is composed of (1) quantitative factors collection, (2) fuzzy IF–THEN rules based model, and (3) decision integration model. The quantitative factors include profit, capacity, due date, and inventory, while the fuzzy IF–THEN rules are summarized from the questionnaire survey for the production experts and learned by a proposed FNN with initial weights generated by real-coded GA. Then, the results from the above two parts are integrated through a feedforward ANN with EBP learning algorithm.
The simulation results showed that the proposed FNN could well learn the fuzzy IF–THEN rules provided by the production experts. In addition, real-coded GA is proved to be better than the binary GA both in speed and accuracy. The real-life problem for an internationally well-known OEM (Original Equipments Manufacturing) company, which manufactures the car lamps for most of car companies, like GM and Ford, illustrates that considering both the quantitative and qualitative factors has more accurate results compared with the results obtained by only considering the quantitative factors. Besides, the proposed system outperforms multiple regression model in precision.
The rest of this paper is organized as follows. Section 2 provides some necessary background information while the proposed system is discussed in Section 3. Section 4 presents the simulation results of FNN, while the evaluation results for real-life problems are summarized in Section 5. Discussion and concluding remarks are made in 6 and 7, respectively.
This study has developed a decision support system for order selection in EC. Considering both the quantitative and qualitative factors it outperforms the case when only quantitative factors are considered. The proposed FNN is able to learning the fuzzy IF–THEN being obtained from the domain experts. In the knowledge-based economics, this method does provide a new tool for capturing the experts' knowledge. The model evaluation results indicate that computational intelligence approach can provide more reliable result than statistical method. In the future, the authors would like to further improve the capability of FNN, like applying the gray theory to improve the training of FNN. In addition, including more qualitative factors may yield a more precise result. The proposed system structure can also be utilized for other applications, like supplier selection for SCM.
