Multiple response optimisations for the development of reduced-fat cake

Abstract

Optimisation is especially important for those professionals involved in the early stages of product development. In this study, multiple response optimisations by using the desirability function (DF) and utilising superimposed contour diagrams were used for the development of low-fat layer cake. The DF is responsible for modelling the multi-response statistical problem. Cakes were prepared with N-Flate formula (8–16%), PaselliSA2 formula (potato maltodextrins (20–40%), monoand diglycerides (MDGs) and propylene glycol monoesters (PGMEs, 2–6%)) and carboxymethyl cellulose (CMC) formula ((1–2%), MDG and PGME (2–6%)). Cake volume, crust and crumb properties were simultaneously optimised. The best factor combinations for each factor were identified to maximise cake volume, crust and crumb properties. The maximised and improved responses were obtained as follows: 9% for N-Flate, 22.5% for Paselli and 1.1% CMC. The composite desirability values of the responses were 0.953, 0.942 and 0.946 for cakes formulated with N-Flate, Paselli and CMC, respectively. It was concluded that response surface methodology and multiple response optimisation could be successfully used to design and optimise low-fat cake formulations with desirable cake properties.