Henry Ford introduced the moving belt assembly line just over a hundred years ago. For the next fifty years the preferred way of organizing manufacture was to have tightly coupled work stations linked by a material handling system. However, gradually it became apparent that neither the moving belt assembly line, or the traditional job shop it replaced, were ideal. Since then there have been major attempts to come up other ways of organizing manufacture. Key to understanding why these developments occurred is to recognize that they attempt to overcome the impact of variability and disturbances on productivity and quality. As each new development has been tried, new insights into the nature of the variability that limits its performance have arisen and this in turn has led to new system designs.
In this talk I will outline these developments and illustrate the way in which formal system models augment understanding of what limits their performance. I begin by considering the advantages of the moving belt line. I address what happens when the line consists of machines rather than people. Next, I consider group technology, extending the line concept to multi product situations. Alternatively, job shops can be improved, leading to flexible manufacturing systems (FMS). But FMS did not live up to their promise and I show why. Quality became a big issue in the 1980s leading to new system designs. However, these revealed the significance of differences between worker performance, leading to new challenges in system design and control.