Formulaware Breakthrough Allows Scientists Access to the Multidimensional World
Dr. M’Hammed Mountassir, Ph.D., V.P. R&D MDM Technologies, Yvon Brousseau, B.Sc, MBA, CTO MDM Technologies
Scientists in industrial research have always recognized that going beyond the 3-D analysis would be the breakthrough required to create globally optimized products, techniques and processes, in the shortest time. Short of this breakthrough, the scientists have tried various methods and technologies in their quest to determine the optimal formula of a product.
DOE methods were selected based on their ability to meet the objectives of the experiment, the number of parameter “Xi” and levels investigated, and the availability of resources. The following represents a few examples of the DOE approach: Randomized block design for comparative objective concludes about one a-priori important parameter and determines whether the parameter is “significant” or not. Full or Fractional Factorial or Plackett Burman design for screening objectives selects or screens out the few important main effects from the many less important “Xi”. Central composite or Box-Behnken design for response surface method (RSM) objective gives an idea of the (local) shape of the response surface properties of each “Yj”. Mixture Design for optimizing responses identifies the “best” proportions of the parameter “Xi” in order to maximize or minimize a response “Yj”. Regression model for an optimal fitting of a regression model objective allows modeling of a response as a mathematical function of a few continuous parameters “Xi” where “good” model factor “Yj” estimates are desired. Taguchi Methods uses orthogonal array designs, and goes beyond fractional factorial by introducing parameter design for increased robustness and tolerance design for quality control. Quantisweb formulaware, in contrast breaks through the 3-D analysis and generates the optimal combination of parameter values “Xi”, simultaneously satisfying all objectives “Yj”.