Designing by self-organization

When designing mechanical systems, the method people normally use is based on what is called reductionism. Reductionism is very simple and powerful, because it is based on the one-to-one correspondence between the required functions and the necessary components of the system. Further, reductionism and the concept of centralized control are naturally suited to each other, because relations between functions and components are clearly seen in centralized control systems. The theme of this book is the theory of design utilizing self-organization, which is the opposite of reductionist design. In designing by self-organization, rather than assigning functions directly to components, relations between components are specified so that the components will organize the whole structure by themselves and consequently let the functions of the whole system emerge. A system built in this way does not have a central component, but instead has decentralized, distributed structure. Although it might seem a very roundabout way of building an entire system, it is a key to progress beyond the limits of reductionist design, by which it is increasingly difficult to meet demands for increasing complexity and scale. Engineering is a discipline which studies methods for creating useful things. The theory of design by self-organization is nothing other than an attempt to establish a new conceptual basis for creating such methods.