This paper describes a computational model that provides planning information useful for scheduling the design process. The model aims to reduce uncertainty in the design process and with it the risk of rework. The view is taken that planning is concerned with choosing between alternative actions and action sequences, but not with resource allocation. The planning model is based on an explicit representation of the state of the design process, the definition of the design capabilities as a pool of tasks, and on the generation and selection of plans by evaluating their reliability. Classical decision theory is used for evaluating the plans: a state-action net is built and analyzed as a Markov decision process. The model produces plans based on qualified task dependencies. These plans can be used as a basis for manual and automated scheduling. In an example industrial case study, a reduction of over 30% in the expected rework was predicted.

Design process planning is essential for business success. However, it is complex for all but the simplest products. There is therefore a need for eflective support for the design process planner to help them visualize and assess alternative design process strategies. This paper describes a new approach to design process planning, known as sigriposting, which has the potential to assist in the characterization, and hence planning, of a particular design process, by identifying a number of possible design routes. The success of signposting is critically dependent upon the planner’s ability CO visualize the emergent process characteristics. A number of route visualization techniques are therefore presented and discussed. The preferred approach is then developed further in an attempt to nzaximize its visualization potential.