Multi-response Robust Design Based On Comprehensive Weight And Cost

With the rapid development of economy,the market has changed greatly,especially in the high demand for product quality.Only if quality can be considered when designing,the problems can be solved fundamentally.Statistics show that,design costs only account for 10%-15% of the total cost,but it determines 70%-80% of it.At the same time,there are 80% quality problems caused by design.Robust design,an important supporting technology in product design,reduces the sensitivity to the noise factor of a product or process by selecting proper level combination of controllable factors.Due to the continuous improvement of customer demand,there are multiple quality characteristics rather than single quality characteristic to be considered when selecting a product.Many quality characteristics may not reach the optimal value at the same time,or even contradictory,so multiple robust design is better for its overall optimization of multiple quality characteristics.Meanwhile,we are chasing the scheme with not only the best quality but also the least cost.In addition,there are also quality characteristics with adjustable values which are called dynamic quality characteristics.The output of a dynamic characteristic is not a fix value but a function of the signal factor,which can be changed to meet the needs of different environments.This paper mainly talks about two issues: the first one was the determination of comprehensive weights which can combine the multiple quality characteristics.In order to do it,the entropy weight was employed to decide the objective weights.If the multiple quality characteristics have the same importance,then the subjective weights would be get by average weight,if not,then expert scoring method would be used to determine the subjective weights,and the scores must meet the requirement of Kendall coefficient test to ensure that the experts can reach a consensus.After that,the comprehensive weights could be get by the combination of objective weights and the subjective weights.The second issue was concentrating on the optimization of robust design when considering the cost.The DEA model was applied,with taking all the combination of different levels of controllable factors as decision-making units,cost as input index and the two series of grey relational degree as the output index,to find the effective DMU.If there were more than one effective DMU,then DEA game across efficiency model was employed for further sorting effective DMUs.The above robust design model can be applied to both static and dynamic quality characteristics.Finally,two examples were analyzed to illustrate the advantage and feasibility of the model.