Page 24 - 6686

х i max = (X i max - X i aver) / h і= +1, (4.5)
х i aver = (X i aver - X i aver) / h і= 0. (4.6)

If each factor varies at two levels (top and low) then for
2
two factors (n = 2) it is needed to conduct two experiments (2 =
4), where 2– number of levels, and n – quantity of factors.
Whith the planning of experiment at three levels (top,
n
middle, low) we get 3 levels and two factors (3 2 = 9
experiments).
After conversion of all factors in conventional scale the
planning matrix is being drafted.
Table 1.2 – Planning matrix for two factors at two levels

Number of Variable of
Levels of factors
an х 0 condition
experiment х 1 х 2 … х k у 1 у 2 … у m
1 +1 +1 +1 у 1,1 у 2,1 … у m,1
2 +1 -1 +1 у 1,2 у 2,2 … у m,2
3 +1 +1 -1 у 1,3 у 2.3 … у m,3
4 +1 -1 -1 у 1,4 у 2,4 … у m,4
Annotation: Х о – column of values of dummy variable.
It was proved that its participation in the planning matrix
gives an opportunity to summarize the calculations of the
coefficients of mathematical model. Experiment begins after a
planning matrix is built.

4.2 Definition of coefficients of linear equation of
regression
After the experiments, according to values of function of
response у і equation of linear model or regression (dependence
of the average value of any parameter on some other
24