用项目管理提高我国芯片开发的效率

Improve Efficiency of Chip R&D Project with Project Management

微电子行业芯片研发中的"摩尔定律"表明科学技术的进步速度已经超过了生产过程中将其转化为生产力和新产品的速度。因此时间短、投入大、技术先进复杂就成为芯片开发项目的最大特征。如何更有效地提高技术转化效率,加快芯片研究速度也成为决定一个公司能否生存的关键。传统的技术革新、改造与新产品开发的管理方式是利用现有的企业资源来整合新产品线,因此必然使新产品研发速度越来越落后于"摩尔定律"下的技术进步速度;而项目管理方式则是根据新产品的技术与功能来整合资源,以有效利用已有资源和引进新资源来整合新的生产线,不仅大大提高了新产品的研发效率,也使新产品研发速度越来越趋近于"摩尔定律"下的技术进步速度。本文将项目管理作为芯片研发过程的最优管理形式;将项目管理的工作分解、计划方法等应用在实际芯片开发过程中;针对芯片开发过程中经常出现的并行任务(任务的耦合性),提出了采用重迭工作法来安排计划等构思。

Moore law in the field of microelectronics has proved that the speed of technological development is much bigger than the transforming speed of changing them into productivities and new productions, which is supposed to be a waste of resource and will lead to a failure in competition with others. This paper testifies how to use project management to resolve this problem. Through combining the theory of modem project management with microelectronics industry, this paper introduces how to make a plan and design a WBS （work breakdown structure） to develop a new chip in reality. To resolve the problem of over-iterations which are always happening in the process of new chips＇ R＆D, a theory of design structure matrix （DSM） has been applied to develop a suitable and economical plan. In addition, this paper clarifies a theory about how to build up a team for a project, control a plan and value a project, etc., which should be attached more attention by project manager.