MV-3010 Tutorial focused in Fstudent

Altair Forum User

Hi gents,

 

I am doind the MV-3000 and MV-3010 tutorials, and I would like to understand what they want to measure with the Sum of squared errors.

Is something I never heard of, and I dont know how important is.

As a definition:  'It is a measure of the discrepancy between the data and an estimation mode'

From the tutorial: 

-	The elements of Vector 1 contain actual data points of the toe curve from the solver run for the nominal configuration.

-	The elements of Vector 2 contain data points from the target curve.

 

I believe that as much room I give to the suspension points to move, as closer I would get to the corner that is defined, right?

 

I imagine that if I set the variables as the suspension points in x,y & z, set the Roll Centre as a response and I define the objective as minimize it I could also have defined my problem.

And from then onwards I could try to fit my toe and camber curves to what I find ideal.

 

Woudl be that the best way of setting up the case?

 

Thanks in advance,

 

Gtn

 

 

 

 

 

 

 

Rahul Rajan_21763

Sum of squared error is one of the common statistical method to determine the error. you can learn about it in any statistics book or refer https://en.wikipedia.org/wiki/Residual_sum_of_squares

 

If you give more room for the suspension points, it help the optimizer to explore those regions but your optimal solution may or may not be in the corner.

 

If you want to optimize Roll centre and Toe, Camber you can try multi-objective Genetic algorithm (MOGA) or Global response Search method (GRSM). 

You can find the details of the these optimization algorithms in HyperStudy documents.