抄録
As the need for the company designers to consider product lifecycle has arose, the number of design criteria and variables has mushroomed. In addition, due to the diversification of users' requests and the rapid changes in fashion, companies must reduce their product development cycle times to an absolute minimum. Consequently, companies now face a two-fold difficulty: design projects that can only be dealt with as large-scale optimization problems, and the need to shorten the development cycle to an absolute minimum. This paper proposes a concurrent criteria-ordering method for large-scale design optimization problems that not only yields an optimum criteria order for evaluation, but also shows how criteria groups can be optimized in parallel. This method consists of two optimizations. First, the input and output data for each criteria will be optimized. Then the order of criteria for evaluation will be optimized according to the optimized input and output data.
