Scalable Platform Robust Reconfiguration Method with Sensitivity Analysis and Fuzzy Clustering
For improving the quick responses and instances quality in platform based product family design, a scalable platform reconfiguration method using sensitivity analysis and fuzzy clustering is proposed with robust design theory. The possible platform constants and scaling variables sets are divided by design variable sensitivity analysis firstly, during which multiple performances are aggregated by preference aggregation to ease the sensitivity for total product. Then the performance preference, robust deviation and constraint violation changes caused by constant sharing are clustered by fuzzy c-means algorithm (FCM) to reasonably plan the constant multiple platforms sharing strategy. A fuzzy percentage index is introduced in FCM to determine the optimal cluster number in fuzzy clustering. Finally the instances deviated from platform are optimized to set scaling variables. The efficiency and effectiveness of proposed method is illustrated by the optimization of scalable capacitor-run single-phase induction motor families, and the results are compared against previous work.
Z. K. Li et al., "Scalable Platform Robust Reconfiguration Method with Sensitivity Analysis and Fuzzy Clustering", Applied Mechanics and Materials, Vols. 52-54, pp. 1026-1031, 2011