A New Mapping Super-Simplex Optimization for Food Product and Process Development

Graphic illustration of response surface is difficult to carry out for cases with more than two factors as the graphs are of higher orders than three dimensions. An approximation was made by dividing the level values of each factor into four groups of equal intervals, and then the response values plotted were linked to each other when all other factor levels were “in the same groups” of experimental conditions, instead of “under the same conditions” in the case of two or three dimensional graphics. When applied to mathematical models, this new mapping simplex optimization significantly (p < 0.01) reduced the number of experiments required for reaching the optimum as compared to the Morgan-Deming simplex optimization. Most experiments on food analysis and processing have been optimized within 25 – 35 iterative experiments depending on the number of factors. Computer-directed optimization in conjunction with the mapping technique followed by a simultaneous shift of factor levels greatly facilitated research supervision.