A PROBABILISTIC-BASED DESIGN MODEL FOR ACHIEVING FLEXIBILITY IN DESIGN

摘要

In the early stages of product development, therntransformation between design requirements and designrnsolutions often involves uncertainties when specifying therndesired target value for the performance expressed in designrnrequirements. Additionally, to provide flexibility for laterrndevelopment, the design solution obtained is desired to be arnrange rather than a single solution. Our primary focus in thisrnpaper is on developing a probabilistic-based design model as arnbasis for providing the flexibility that allows designs to bernreadily adapted to changing conditions. This is obtained byrndeveloping a range of design solutions which meet a ranged setrnof design requirements. Meanwhile, designers are allowed tornspecify the varying degree of desirability of a ranged set ofrndesign requirements based on their preferences. The DesignrnPreference Index (DPI) is introduced as a design metric tornmeasure the goodness of flexible designs. Providing thernfoundation to our work are the probabilistic representations ofrndesign performance, the application of robust design concept,rnand the utilization of the compromise Decision SupportrnProblem (DSP) as a multi-objective decision model. A two-barrnstructural design is used as an example to demonstrate ourrnapproach. Our focus in this paper is on introducing thernprobabilistic-based design model and not on the results of thernexample problem, per se.