With JMP, i have made a D-optimal Design for an experiment with 14 runs in the presence of such a time trend, for a special cubic scheffé model (third-order model) but without interaction between the time trend an the mixture ingredient factors.
Based on the insights I acquired, I have to construct a second trend-robust D-optimal design and explain the principle I applied.
I suggest that I have to find a trend-robust D-optimal design that better fits the data, but no idea how to do that.
Can anyone help me?