A quantitative model for the design of nutritious and acceptable foods : a thesis presented in partial fulfilment of the requirements for the degree of Ph. D. in Product Development at Massey University
A quantitative model was developed for the systematic selection of raw materials for nutritious and acceptable Thai foods. The basis for the quantitative model was linear programming. Unlike most other applications of this technique which considered only nutritional needs, this model also took into account the consumer's requirements for an acceptable food. The development of the model formed part of a research programme in the Food Technology Department of Massey University which is investigating quantitative product development techniques. The linear programming model was used firstly to select raw material mixes which would satisfy the daily nutritional requirements of the Thai people. Selection was made from a list of 151 available raw materials - 144 indigenous Thai raw materials and 7 New Zealand dairy products. The requirements for 26 nutrients were satisfied. These included protein, fat, calories, fibre, 3 minerals, 9 vitamins and 10 essential amino acids. The lack of specific upper limits on most of the nutrients resulted in solutions with a gross nutritional imbalance. Recent investigations indicate that such imbalances may be detrimental to human health and it is suggested that a more satisfactory diet is one where all nutrients are balanced, at or near their lower requirement levels. Considerable problems exist in achieving a balanced diet using range constraints in the linear programme due to the probability of solution infeasibilities. Goal programming, an extension of linear programming, has been used in other fields of research to minimize the deviation of solution variables from specific goals. This technique showed potential in attaining a balanced nutritional diet where the goals represented the requirements for specific nutrients. A goal programming model was devised which firstly achieved a balance of essential amino acids as close to that of egg protein as possible. Secondly, a solution was obtained where all 26 nutrients were at the 'optimum' balance. The achievement of a balanced nutritional diet resulted in a large increase in cost and indicated the importance of careful definition of the requirements for nutritional balance in future research where cost minimization is a priority. The raw materials selected by both goal and linear programming were totally unacceptable as ingredients in a Thai food dish without extensive processing to change both flavour and texture of the mixture. It was more logical to provide a procedure for raw material selection on the basis of their combined acceptability in a Thai food dish. Nonmetric multidimensional scaling was used to derive a 3 dimensional configuration of 40 raw materials from consumer information on the use of these raw materials in Thai food dishes. The axes of this space represented the dominant properties of raw materials in determining food dish acceptability. An ideal point was located in this space. This point was defined as the 'optimum' combination of raw material properties required in a Thai food dish. Nonmetric multidimensional scaling provided the basis for derivation of metric scale values for the 40 raw materials and the ideal point. These values were used to derive a linear function relating raw materials to food dish acceptability. This function was used in the linear programming model together with nutritional constraints to provide a systematic method of raw material selection for nutritious and acceptable Thai food dishes. The raw material mixes selected by this model showed a marked improvement over those chosen by the linear programme subject to only nutritional constraints.