Turning 1D MVar into a diagonal matrix

回答済み

Gabriel Agostini

• 2024年04月30日 22:39

Hi,

I wonder if GUROBI has a function that creates a diagonal matrix out of a 1-D MVar. I essentially have as optimization variables these "row-wise scaling factors" k, which I need to multiply by a given matrix to obtain a matrix M. So far I am trying to do:

#N_1 : int
#E   : N_1 x N_1 2d csr sparse matrix

#Define our variables (N_1 scaling factors):
K = model.addMVar(shape=N_1, lb=0)

#Define the transformation M which is the row-scaled version of E:
M = diag(K) @ E

This last line doesn't really work. I also tried creating a sparse matrix by accessing K.X for example but the value wasn't available at the first pass. I also cannot use for loops as my N_1 value is very large.

Is there a way to diagonalize my 1D MVar K, or maybe is there a way to multiply a matrix row-wise by an 1D MVar?

Thanks!

関連するエンティティ

• Matrix API

0

• 

  Riley Clement

  Gurobi Staff

  • 2024年05月01日 00:53

  Hi Gabriel,

  I think

  E*K.reshape(N_1,1)

  is what you are looking for.

  - Riley

   

  0
• 

  Gabriel Agostini

  • 2024年05月01日 18:28

  

  

  Thanks Riley Clement ! That works

  0

サインインしてコメントを残してください。