My problem is related to the design evaluation of my set of experiments. I am trying to design a mixture of 4 materials, each having certain minimum and maximum dosage constraints. To come up with a valid design I am following a D-optimal mixture design, using the software of Design-Expert, and the objective is to obtain an optimal mixture proportion of these raw materials, in terms of certain properties. However, as I enter the constraints and get the design and proceed to the evaluation of the design, the important parameters of standard errors, variance inflation factor, R-Squared and the powers are really poor compared to their criteria (e.g. for some terms the variance inflation factor is more than 4000). I considered some transformations, but because of the constraints I could not find and apply a proper one. Also, the model reduction has not been effective. My question is, being in this stage that I have not yet performed my experiments ( I have not entered any data and this is just the evaluation of the design of experiments), does this mean that the design is poor and ineffective? In another way, does this mean that even if I perform the tests, the data will not be efficiently modelled and analysed? Any ideas would be a great help. Thanks!