The elements of the matrices are randomly generated.
Some tests:
n = 150 time = 4877 milliseconds n = 140 time = 4132 milliseconds n = 130 time = 3197 milliseconds n = 120 time = 2513 milliseconds n = 110 time = 1926 milliseconds n = 100 time = 1612 milliseconds n = 90 time = 1073 milliseconds n = 80 time = 772 milliseconds n = 70 time = 515 milliseconds n = 60 time = 327 milliseconds n = 50 time = 193 milliseconds n = 40 time = 119 milliseconds n = 30 time = 41 milliseconds n = 20 time = 30 milliseconds n = 10 time = 9 milliseconds
I can graphically represent the data (see the blue dots):
n: 0,10,20,30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150 t: 0, 9,30,41,119,193,327,515,772,1073,1612,1926,2513,3197,4132,4877
I sense that the trend could be quadratic. I check this by calculating the first and second differences (see).
The second differences oscillate around 0: I deduce that the trend is approximately quadratic. I use quadratic regression (see):
I superimpose the graph on the points and I get the red curve in the previous diagram.