英文摘要 |
The planning and design (P&D) phase is most critical to project success in new product development (NPD). When selecting an appropriate P&D alternative which involves a variety of considerations such as corporate strategy, finance, marketing, technology, production, customer need, product aesthetics, and so forth, most company's high-level management lack the ability to evaluate the candidates effectively. Substandard NPD project is often a direct result of inadequate decision-maker selection. In this paper, we propose a concurrent SMART^2 evaluation model of NPD design alternatives using the bi-directional Analytic Network Process (ANP). It is utilized to determine the local weights of these candidates from both planning and design dimensions. Following the above evaluation model, a Gradient Sensitivity Analysis (GSA) is employed to calculate the overall weights and a mathematical Zero-One Goal Programming (ZOGP) is built to incorporate multiple objectives in order to reach an optimum solution. A NPD design case is presented to illustrate the overall decision procedure and to examine the effectiveness of our proposed model. The result shows that the local weights of alternatives in the planning aspect are C (0.365) > B (0.273) > A (0.257) > D (0.105) while those in the design aspect are A (0.299) > C (0.237) > B (0.283) > D (0.181). Furthermore, the priority of alternatives in the largest of the three decision areas (C > A > B > D) identified by GSA is around 66.5% (20.5%-87%). Finally, a 50-50 proportion in planning and design dimensions is applied as a simulation point to demonstrate the final optimal solution-design alternatives A and C. |