Hybrids of Probability-Based Multi-objective Optimization with Experimental Design Methodologies
摘要
It describes the hybrids of probability-based multi-objective optimization with experiment design methodologies, including orthogonal experimental design, response surface design, and uniform experimental design. The total preferable probability of a candidate alternative is the unique decisive index for the alternative selection or optimization quantitatively. The optimization of multi-objective orthogonal experiment design is conducted by using range analysis to the total preferable probabilities comprehensively; the optimizations of multi-objective of response surface design and uniform experiment design are to get maximum of the total preferable probability integrally. Some application examples in materials selection are given.