最佳化實驗設計在統計實務應用中，針對不對稱的實驗區域設計方法有其存在之必要性。本研究使用修正的質群演算法於一連串D 最佳化設計問題，且在搜尋最佳解中找出適當方法平衡質群演算法於速度更新的搜尋能力。利用保留區域可行解重新設定不可行解在實驗區域邊界上的方式，進而處理混合實驗與複雜限制式的問題，實驗結果驗證本方法可獲得良好的績效。

Alphabetic designs are applied extensively to the engineering design problems as the standard experimental designs cannot be directly used or some design variable is restricted due to practical limitation. In this situation, computer-generated designs are a widely accepted alternative to deal with an irregular region of experimentation, a nonstandard model or specific design criteria. The most popular one is D-optimal designs which would minimize the volume of the joint confidence region on the vector of regression coefficients. This paper presents a new optimization technique to generate D-optimal designs using the modified particle swarm optimization (termed MPSO).
The prominent merit of the proposed method is the ability to more likely reach a global optimal set of design points than using exiting techniques as a mixture design is solved. The experimental results obtained from running a set of test problems are used to illustrate MPSO.