基于神经网络和遗传算法的变型设计研究

Case study of variant design based on neural network and generic algorithm

随着社会生产力高速发展,市场中的产品表现出以下趋势:功能标准化、造型多样化。在这两种趋势下,变型设计对产品的市场效益逐渐起主导作用。变型设计需要少数用户参与交流,所获取的用户知识有限,无法推广至整个市场。该文提出的变型设计,先用神经网络学习整个市场范围内的用户知识,然后用遗传算法推断与其最为匹配的设计知识。最后,以手机设计为例,探讨了该方法的应用。

With the high - speed development of social productivity, the products over market have shown two trends： function being standardized and form being diversified. In that case, variant design begins playing a dominant role in product＇ s benefit. However, variant design needs commucination between stylist and few users, in which the user knowledge acquired is too limited to be extended towards the whole market. The variant design this paper proposed firstly learns the user knowledge over the whole market by neural network, then infers the design knowledge most matching the user knowledge. Finally, this paper discusses the application of this method in mobilephone design.