A new approach for predicting the knit global quality by using the desirability function and neural networks

Quality is a very important notion, which depends on every person’s taste, on the conditions of use, and the requirement degree of every consumer. The global quality of the knit is a multicriteria phenomenon that requires the satisfaction of several properties at the same time; among these properties are the properties of comfort, permanency, usefulness, and the mechanical resistance properties. In this paper, we used the individual desirability functions to evaluate the satisfaction degree of each property defining the quality, and we also used the global desirability function to foresee the global quality by according relative weight to each individual satisfaction degree. For each consumer, these relative weights depend on the importance of each property. We use the neural network to foresee the global quality of the knitted fabric from the fiber, yarn, and knitting parameters. This neural network is optimized by using an algorithm, which is modified when the definition of knit quality is changed. Finally, this forecasting can help the knitter to estimate if the knitted fabric that he/she will produce will match the consumer's taste. Thus, he/she can adjust the knitting production parameter to reach the desired quality to satisfy his/her consumer.