Hi, We trying to solve the next optimization problem? We have: Pij, i=1÷3, j=1÷30, Pij are positive Bi, i=1÷3, integer positiveConstraints: The searching result is matrix of 3 x 30 of binary values Xij with next conditions: For each j =1÷30, Sum (by index of i=1÷3)Xij=1 For each i =1÷3, Sum by index of (j=1÷3o)Xij≤Bi Objective: Optimize: Maximize (Sum (by index of i=1÷3) Sum (by index of j=1÷30) Pij *Xij)Is there another predefined subroutine in SAS/IML which can be used to solve this problem? We haven’t the SAS/OR. The suggested solution from Rick_SAS was to convert this to linear optimization programing. You can see here the solution. Linear optimization in SAS/IML In real situation we for j index we have 30000, not only 30. The matrix became too large. It is not possible to solve it by SAS/IML. The limitation there is 200 constraints, which is too few. We haven’t SAS/OR yet. My question concerns the limitation of constraints and the size of matrix in SAS/OR? Is it possible to work with matrix size of 90000 x 30000. How many constraints there can have? Is it possible to solve this task with another procedure, not with linear programing in SAS/OR? Thanks, Boriana
... View more